CN112488831B

CN112488831B - Block chain network transaction method and device, storage medium and electronic equipment

Info

Publication number: CN112488831B
Application number: CN202011314898.1A
Authority: CN
Inventors: 牟童; 王诗鈞; 徐石成; 何光宇
Original assignee: Neusoft Corp
Current assignee: Neusoft Corp
Priority date: 2020-11-20
Filing date: 2020-11-20
Publication date: 2024-08-23
Anticipated expiration: 2040-11-20
Also published as: CN112488831A

Abstract

The disclosure relates to a blockchain network transaction method, a device, a storage medium and electronic equipment, wherein the method comprises the following steps: responding to a transaction request acquired by a target node, acquiring state data corresponding to target historical transaction of the target node; inputting the state data into an environment prediction model to obtain a target type label of the state data output by the environment prediction model; determining a target transaction success rate of the transaction request according to the proportion of training samples marked with the target type labels and having the transaction state of success; and determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold.

Description

Block chain network transaction method and device, storage medium and electronic equipment

Technical Field

The present disclosure relates to the field of blockchain technologies, and in particular, to a blockchain network transaction method, a device, a storage medium, and an electronic apparatus.

Background

Blockchains are a technique that enables collective maintenance of a reliable database by means of decentralization and de-trust. The method can store the transactions occurring in a period of time in blocks as units, and connect the blocks in time sequence by a cryptography algorithm to form a data structure similar to a chain. The blockchain technology has the characteristics of distributed account book, decentralization, non-falsification and the like, and has a high application prospect in various aspects.

In the related scenario, when the blockchain node performs the cross-chain transaction, a transaction failure phenomenon may occur, so that the transaction rollback is caused, and the performance of the whole cross-chain system is further reduced.

Disclosure of Invention

The disclosure provides a blockchain network transaction method, a blockchain network transaction device, a storage medium and an electronic device, so as to solve the above related technical problems.

To achieve the above object, according to a first aspect of embodiments of the present disclosure, there is provided a blockchain network transaction method, including:

Responding to a transaction request acquired by a target node, acquiring state data corresponding to target historical transactions of the target node, wherein the target historical transactions are historical transactions generated by the target node within a preset time before the current moment, or the target historical transactions are preset number of historical transactions which are generated by the target node recently before the transaction request is acquired;

Inputting the state data into an environment prediction model to obtain a target type label of the state data output by the environment prediction model, wherein training samples of the environment prediction model are obtained based on the state data of historical transactions generated in a blockchain network, and each training sample is marked with a type label;

Determining a target transaction success rate of the transaction request according to the proportion of training samples marked with the target type labels and having the transaction state of success;

And determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold.

Optionally, the status data includes transaction status data of the target historical transaction, and the environmental prediction model includes a transaction status prediction model, and accordingly:

the step of inputting the state data into an environment prediction model to obtain a target type tag of the state data output by the environment prediction model comprises the following steps:

Inputting transaction state data of all target historical transactions as one input quantity of the transaction state prediction model into the transaction state prediction model to obtain a first target type tag which is output by the transaction state prediction model and represents the transaction state type, wherein training samples of the transaction state prediction model are in one-to-one correspondence with historical transaction time periods of the blockchain network, and transaction state data corresponding to all historical transactions of the same blockchain node in one historical time period are used as one training sample;

the determining the target transaction success rate of the transaction request according to the training samples marked with the target type tag and the transaction state being the successful state, wherein the proportion of the training samples marked with the target type tag comprises the following steps:

And determining a first transaction success rate of the transaction request according to the training samples marked with the first target type tag and the transaction state being the success state, wherein the first transaction success rate of the transaction request is determined according to the proportion of the training samples marked with the first target type tag, and the target transaction success rate comprises the first transaction success rate.

Optionally, the state data includes system state data for characterizing a system state at an occurrence time of the target historical transaction, and the environmental prediction model includes a system state cluster model, and accordingly:

Inputting system state data of all target historical transactions into the system state clustering model in a mode that the system state data of one target historical transaction is used as an input quantity, obtaining a second target type label which is output by the system state clustering model and represents the system state type of each target historical transaction, wherein a training sample of the system state clustering model is the system state data corresponding to each historical transaction of the blockchain network, the system state clustering model clusters into a plurality of clusters aiming at all training samples in the training process, and the second target type label is a target cluster to which the system state data representing the target historical transaction belongs;

And determining a second transaction success rate of the transaction request according to the training samples marked with the second target type tag and the transaction state being the successful state, wherein the second transaction success rate of the transaction request is determined according to the proportion of the training samples marked with the second target type tag, and the target transaction success rate comprises the second transaction success rate.

Optionally, the determining, according to the training samples marked with the second target type tag and the transaction state is the successful state, the proportion of the training samples marked with the second target type tag in all the training samples includes:

determining, for each of the second target type tags, a first duty cycle of training samples marked with such second target type tag in all training samples;

For each second target type tag, determining a second duty ratio of system state data belonging to the second target type tag in state data corresponding to all target historical transactions;

determining the weight of the second target type tag corresponding to the system state data of each target historical transaction according to the first duty ratio and the second duty ratio of each second target type tag;

Weighting and summing the weights of the second target type labels corresponding to the system state data of each target historical transaction and the target proportion corresponding to the second target type labels marked with the system state data to obtain the second transaction success rate;

The target proportion is a training sample marked with the second target type label and the transaction state is a successful state, and the proportion of the training sample marked with the second target type label is occupied in all training samples.

Optionally, the state data comprises system state data of the target historical transaction, and the environmental prediction model comprises a system state prediction model, and accordingly:

Taking system state data of all target historical transactions as an input quantity of the system state prediction model, inputting the system state prediction model to obtain a third target type label which is output by the system state prediction model and represents the system state type, wherein training samples of the system state prediction model are in one-to-one correspondence with historical transaction time periods of the blockchain network, and system state data corresponding to all historical transactions of the same blockchain node in one historical time period is taken as a training sample;

And determining a third transaction success rate of the transaction request according to the training samples marked with the third target type tag and the transaction state being the success state, wherein the third transaction success rate of the transaction request is determined according to the proportion of the training samples marked with the third target type tag, and the target transaction success rate comprises the third transaction success rate.

Optionally, the state data includes transaction state data of the target historical transaction, and the environmental prediction model includes a transaction state cluster model, corresponding to:

inputting transaction state data of all the target historical transactions into the transaction state clustering model in a mode that transaction state data of one item of target historical transaction is used as an input quantity, obtaining a fourth target type label which is output by the transaction state clustering model and represents the transaction state type of each item of target historical transaction, wherein a training sample of the transaction state clustering model is transaction state data corresponding to each item of historical transaction of the blockchain network, the transaction state clustering model clusters into a plurality of clusters aiming at all training samples in the training process, and the fourth target type label is a target cluster to which the transaction state data representing the target historical transaction belongs;

And determining a fourth transaction success rate of the transaction request according to the training samples marked with the fourth target type tag and the transaction state being the success state, wherein the fourth transaction success rate of the transaction request is determined according to the proportion of the training samples marked with the fourth target type tag, and the target transaction success rate comprises the fourth transaction success rate.

Optionally, the state data includes transaction state data of the target historical transaction and system state data, the system state data is used for representing a system state at the occurrence time of the target historical transaction, and the environment prediction model includes a transaction state prediction model, a system state clustering model, a transaction state clustering model and a system state prediction model; accordingly:

the target transaction success rate comprises:

a first transaction success rate obtained based on the transaction state data and the transaction state prediction model;

a second transaction success rate obtained based on the system state data and the system state clustering model;

A third transaction success rate obtained based on the system state data and the system state prediction model;

a fourth transaction success rate obtained based on the transaction state data and the transaction state clustering model;

the determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold value comprises the following steps:

The first transaction success rate, the second transaction success rate, the third transaction success rate and the third transaction success rate are weighted and summed to obtain a sum value;

Executing the transaction operation corresponding to the transaction request under the condition that the sum value is larger than the success rate threshold value;

and under the condition that the sum value is smaller than or equal to the success rate threshold value, the transaction request is cancelled.

Optionally, the transaction status data includes one or more of transaction concurrency number, transaction response time, transaction throughput, transaction timeout setting time, chunk out time setting information, chunk size setting information, last transaction result identification, and current transaction result identification.

Optionally, the system state data includes one or more of a central processing unit CPU occupancy rate, a network occupancy bandwidth, a memory occupancy rate, a disk usage rate, a CPU occupancy rate change, a network occupancy bandwidth change rate, a memory occupancy rate change, and a disk usage rate change.

Optionally, the target cluster number of the system state cluster model or the transaction state cluster model is obtained by the following method:

setting a plurality of candidate cluster numbers;

Clustering the data to be clustered according to the candidate cluster number aiming at each candidate cluster number until the cluster center converges;

Calculating inter-class distance values of a plurality of cluster categories corresponding to each candidate cluster number and intra-class distance values of data in each cluster category;

And taking the candidate cluster number corresponding to the maximum value in the difference value between the inter-class distance values and the intra-class distance values as the target cluster number.

Optionally, the transaction request is a cross-link transaction request, and accordingly, the target historical transaction is a historical cross-link transaction generated by the target node within a preset time before the current moment, or the target historical transaction is a preset number of historical cross-link transactions generated by the target node recently before the transaction request is acquired, and training samples of the environment prediction model are obtained based on state data of the historical cross-link transactions generated in the blockchain network.

According to a second aspect of embodiments of the present disclosure, there is provided a blockchain network transaction device, comprising:

The first acquisition module is used for responding to a transaction request acquired by a target node, acquiring state data corresponding to target historical transactions of the target node, wherein the target historical transactions are historical transactions generated by the target node within a preset time before the current moment, or the target historical transactions are preset number of historical transactions which are generated by the target node most recently before the transaction request is acquired;

the first input module is used for inputting the state data into an environment prediction model to obtain a target type label of the state data output by the environment prediction model, training samples of the environment prediction model are obtained based on the state data of historical transactions generated in a blockchain network, and each training sample is marked with a type label;

the first determining module is used for determining the target transaction success rate of the transaction request according to the proportion of training samples marked with the target type tag and the transaction state being the successful state in all the training samples marked with the target type tag;

and the second determining module is used for determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold value.

According to a third aspect of embodiments of the present disclosure, there is provided a computer readable storage medium having stored thereon a computer program which, when executed by a processor, implements the steps of the method of any of the first aspects described above.

According to a fourth aspect of embodiments of the present disclosure, there is provided an electronic device, comprising:

A memory having a computer program stored thereon;

A processor for executing the computer program in the memory to implement the steps of the method of any of the above first aspects.

In the above technical solution, when a target node in a blockchain network acquires a transaction request, the target node may acquire state data corresponding to a target historical transaction of the target node. In this way, the target type label corresponding to the state data can be obtained by inputting the state data into an environment prediction model, and the target transaction success rate of the transaction request is determined according to the proportion of training samples marked with the target type label and the transaction state being the successful state in all the training samples marked with the target type label. That is, the technical scheme can predict the success rate of the transaction before the transaction is performed, and further can cancel the transaction with lower success rate. Therefore, by adopting the technical scheme, the probability of successful transaction can be improved, and the influence of transaction failure on network performance can be reduced.

Additional features and advantages of the present disclosure will be set forth in the detailed description which follows.

Drawings

The accompanying drawings are included to provide a further understanding of the disclosure, and are incorporated in and constitute a part of this specification, illustrate the disclosure and together with the description serve to explain, but do not limit the disclosure. In the drawings:

Fig. 1 is a flow chart of a blockchain network transaction method shown in an exemplary embodiment of the present disclosure.

Fig. 2 is a flow chart of a blockchain network transaction method shown in an exemplary embodiment of the present disclosure.

Fig. 3 is a flow chart of a blockchain network transaction method shown in an exemplary embodiment of the present disclosure.

Fig. 4 is a flowchart illustrating a blockchain network transaction method in accordance with an exemplary embodiment of the present disclosure.

Fig. 5 is a flow chart of a blockchain network transaction method shown in an exemplary embodiment of the present disclosure.

FIG. 6 is a flow chart of determining the number of clusters of a cluster model, as shown in an exemplary embodiment of the present disclosure.

Fig. 7 is a block diagram of a blockchain network transaction device in accordance with an exemplary embodiment of the present disclosure.

Fig. 8 is a block diagram of an electronic device, as shown in an exemplary embodiment of the present disclosure.

Detailed Description

Specific embodiments of the present disclosure are described in detail below with reference to the accompanying drawings. It should be understood that the detailed description and specific examples, while indicating and illustrating the disclosure, are not intended to limit the disclosure.

Before describing the blockchain network transaction method, device, storage medium and electronic equipment of the present disclosure, application scenarios of embodiments provided by the present disclosure are first described, where the embodiments provided by the present disclosure may be applied to transaction scenarios of blockchain nodes, for example, transactions in the same blockchain network may be cross-chain transactions performed between nodes belonging to different blockchain networks.

To this end, the present disclosure provides a blockchain network transaction method, and fig. 1 is a flowchart of a blockchain network transaction method according to an exemplary embodiment of the present disclosure, the method including:

in step S11, in response to the target node acquiring the transaction request, state data corresponding to a target historical transaction of the target node is acquired.

In particular, the blockchain network transaction method may be applied to a transaction recipient, i.e. the target node may act as a transaction recipient. After receiving the transaction request, the target node can acquire state data corresponding to target historical transaction of the target node.

Regarding the target historical transaction, the target historical transaction may be a historical transaction generated by the target node within a preset time period before the current time, and the preset time period may be, for example, one hour, one day, or the like. In some embodiments, the target historical transaction may also be a preset number of historical transactions that were last generated by the target node before the transaction request was obtained, such as 10 last generated transactions, 15 transactions, and so on.

In S12, the state data is input to an environmental prediction model, and a target type tag of the state data output by the environmental prediction model is obtained.

Training samples of the environmental prediction model may be derived, for example, based on state data of historical transactions generated in a blockchain network, it being understood that the composition of the training samples may be consistent with the state data input to the environmental prediction model. For example, if the state data input to the environment prediction model is state data corresponding to 10 historical exchanges generated most recently in history, when training the environment state model, the historical exchanges in the blockchain network may be arranged according to time sequence, and the state data corresponding to 10 adjacent historical exchanges may be used as a training sample.

In addition, each of the training samples may also be labeled with a type tag. In this way, by comparing the type label with the prediction label output by the environmental prediction model, the loss value of the environmental prediction model can be determined, so that model parameters can be adjusted according to the loss value, and the model training effect is achieved.

In this way, the state data can be input into an environment prediction model, and a target type label of the state data output by the environment prediction model is obtained.

In step S13, a target transaction success rate of the transaction request is determined according to the proportion of training samples marked with the target type tag and having a successful transaction state in all training samples marked with the target type tag.

It should be appreciated that since the training samples are derived based on the status data corresponding to the historical transaction, each training sample may correspond to a true transaction outcome, such as a transaction success or a transaction failure. Thus, after the target type tag corresponding to the target historical transaction is obtained through the environment prediction model, the transaction success rate corresponding to the transaction request can be determined based on the historical data. For example, the training samples marked with the target type tag and the transaction state is the successful state, and the proportion of the training samples marked with the target type tag is used as the target transaction success rate of the transaction request.

In step S14, it is determined whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold.

For example, in some implementations, the target transaction success rate is greater than a success rate threshold, in which case the transaction may be performed. In other implementations, the target transaction success rate may also be less than the success rate threshold, in which case the transaction request may be revoked/denied. Further, the success rate threshold may be set according to application requirements, such as 60%,80%, and so on.

It should be noted that, in the above embodiments, the method is described by taking an example that the method is applied to a transaction receiver, but in some implementation scenarios, the method may also be applied to an initiator, a relay, and the like of a transaction. Meanwhile, the initiator and the receiver of the transaction can be in the same blockchain network and can belong to different blockchain networks.

For example, in one possible implementation, the transaction request is a cross-link transaction request, and accordingly, the target historical transaction is a historical cross-link transaction generated by the target node within a preset time period before the current moment, or the target historical transaction is a preset number of historical cross-link transactions generated by the target node recently before the transaction request is acquired, and training samples of the environment prediction model are obtained based on state data of the historical cross-link transactions generated in the blockchain network. That is, the above technical solution can predict the success rate of the cross-link transaction, thereby reducing the influence of the phenomenon of cross-link transaction failure on the system performance.

In addition, in some implementation scenarios, the status data may further include status data of a target historical transaction corresponding to each transaction party. For example, when the method is executed on a transaction receiver, the transaction receiver not only can determine the target transaction success rate according to the state data corresponding to the target historical transaction of the own node, but also can acquire the state data corresponding to the target historical transaction of the transaction initiator, and determine another target transaction success rate through the blockchain network transaction method provided by the disclosure. In this way, the final success rate of the transaction can be determined by integrating the target transaction success rates of the transaction parties respectively, and further, the execution strategy of the transaction can be determined. Of course, in some implementations, the transaction party may also include a relay party to the transaction, which is not limited by the present disclosure.

In one possible implementation, the status data includes transaction status data of the target historical transaction, and the environmental prediction model includes a transaction status prediction model. Referring to the flowchart of a blockchain network transaction method shown in fig. 2, the method, based on fig. 1, inputs the state data into an environment prediction model to obtain a target type tag of the state data output by the environment prediction model, includes:

s121, taking the transaction state data of all the target historical transactions as one input quantity of the transaction state prediction model, inputting the transaction state prediction model, and obtaining a first target type label which is output by the transaction state prediction model and represents the transaction state type.

The transaction status data may represent a transaction status corresponding to each of the target historical transactions, and may include one or more of a transaction concurrency number, a transaction response time, a transaction throughput, a transaction timeout setting time, a block out time setting information, a block size setting information, a last transaction result identifier (e.g., a success identifier or a failure identifier), and a current transaction result identifier (e.g., a success identifier or a failure identifier).

The training samples of the transaction state prediction model are in one-to-one correspondence with historical transaction time periods of the blockchain network, and transaction state data corresponding to all historical transactions of the same blockchain node in one historical time period is used as one training sample.

For example, the transaction state prediction model may be constructed based on an LSTM (Long Short-Term Memory network), for example. Wherein the time step of the LSTM may be determined based on a transaction state of a blockchain network, for example. For example, a time t seconds may be chosen and the average number of transactions generated within t seconds is calculated as the time step NT of the LSTM. In this way, training samples may be determined based on the time step NT and transaction state data corresponding to historical transactions of the blockchain node.

For example, the training samples may be determined based on transaction state data corresponding to a plurality of historical exchanges of the same node. Taking eight characteristics of the transaction state data including a transaction concurrency number T _c, a transaction response time T _r, a transaction throughput T _tps, a transaction timeout setting time T _o, a block out time setting information B _t, a block size setting information B _s, a last transaction result identifier next and a current transaction result identifier current as an example, if the node includes 100 transaction state data corresponding to 100 historical transactions, and the time step TN is 20, a training sample [5, 20,8] can be obtained. Where 20 is the time step, 5 is the number of samples, and 8 is the sample feature number ({ T _c,T_r,T_tps,T_o,B_t,B_s, next, current }) corresponding to the transaction status data.

Of course, in some implementations, different training samples may be generated from the same data set based on the selected differences in starting time points. For example, for the transaction state data 1-11 arranged in time series, in the case where the time step is 10, the transaction state data 1-10 may be regarded as one training sample, or the transaction state data 2-11 may be regarded as one training sample.

In some embodiments, the LSTM may include, for example, NT x 32 x 64 x 128 x 256 neurons. For the LSTM unit of the hidden layer, a corresponding weight value can be set for the LSTM unit.

For example, the hidden state scores at the t-th time point and all time points before can be calculated,

e_t＝W^Th_t

Wherein, the time points correspond to transactions, W ^T＝(h₁,h₂...,h_t,...,h_NT)^T, t ε {1,2, …, NT }, NT is the total number of time points, W ^T is the state of all time points, and h _t is the state of time point t.

Further, the hidden state score of each time point may be normalized based on a softmax function, to obtain an attention value of the hidden state assignment of the t-th time point relative to the hidden state assignment of the s-th time point as follows:

Where e _ts is the t-th element in e _t, s.epsilon. {1,2, …, NT }. In this way, it can be determined that the hidden layer output at the t-th time point under the Self-Attention mechanism (Self-Attention) is

After constructing the LSTM, the determined training sample data may be input to the LSTM to obtain a predicted result of the LSTM. In the above example, if the training samples input to the transaction state prediction model are transaction state data 1 to 10, the type tag of the transaction state data 11 may be used as the type tag of the training samples. In this way, the correctness of the prediction result of the transaction state prediction model can be judged by comparing the prediction label output by the transaction state model with the type label of the training sample, and further relevant network parameters of the LSTM are updated until the model converges, so that training is completed.

Still referring to fig. 1 and 2, in this case, the determining, according to the training samples marked with the target type tag and the transaction status being the successful status, the target transaction success rate of the transaction request according to the proportion of the training samples marked with the target type tag, includes:

S131, determining a first transaction success rate of the transaction request according to the proportion of training samples marked with the first target type tag and the transaction state being the success state in all the training samples marked with the first target type tag, wherein the target transaction success rate comprises the first transaction success rate.

For example, if the first target type label output by the transaction state prediction model is a, and 10000 samples of the label a are training samples of the transaction state prediction model, and 7000 samples of successful transaction are training samples, it may be determined that the transaction success rate corresponding to the first target type label a is 70%, that is, the first transaction success rate corresponding to the transaction request is 70%.

In this case, whether to execute the transaction operation corresponding to the transaction request may be determined based on the magnitude relation between the first transaction success rate and the success rate threshold, and the specific policy is described with reference to the embodiment of step S14, which is not described in detail herein.

According to the technical scheme, the success rate of the transaction corresponding to the transaction request can be predicted based on the transaction state data of the target historical transaction, so that the occurrence frequency of the transaction failure phenomenon can be reduced.

Regarding the type labels of the training samples, in some embodiments, the determination may be based on a clustering approach. For example, all training samples may be clustered by a clustering model, resulting in a plurality of clusters, and accordingly, the type labels of the training samples may correspond to any of the plurality of clusters.

In other embodiments, the state data includes system state data for characterizing a system state at a time of occurrence of the target historical transaction, and the environmental prediction model includes a system state cluster model. In this case, the success rate of the transaction may also be determined based on a clustering model. For example, referring to the flowchart of a blockchain network transaction method shown in fig. 3, the method, based on fig. 1, inputs the state data into an environment prediction model to obtain a target type tag of the state data output by the environment prediction model, includes:

s122, inputting the system state data of all the target historical transactions into the system state clustering model in a mode that one item of system state data of the target historical transactions is taken as one input quantity, and obtaining a second target type label which is output by the system state clustering model and represents the system state type of each target historical transaction.

The system state data comprises one or more of CPU occupancy rate, network occupied bandwidth, memory occupancy rate, disk usage rate, CPU occupancy change rate, network occupied bandwidth change rate, memory occupancy change rate and disk usage change rate of a Central Processing Unit (CPU). The CPU occupancy rate, the network occupancy bandwidth rate, the memory occupancy rate, and the disk usage rate may be determined, for example, according to the CPU occupancy rate, the network occupancy bandwidth, the memory occupancy rate, and the disk usage rate at the time corresponding to the adjacent transaction. In some scenarios, the average CPU occupancy rate, the average network occupancy bandwidth rate, the average memory occupancy rate, and the average disk usage rate may also be selected over a threshold time.

The training samples of the system state clustering model are system state data corresponding to each historical transaction of the blockchain network, the system state clustering model is clustered into a plurality of class clusters aiming at all training samples in the training process, and the second target type label is a target class cluster to which the system state data representing the target historical transaction belongs.

For example, the number K of clusters may be set for all training samples, and K clusters may be obtained finally by selecting a cluster center and updating the cluster center. In this way, each cluster can be treated as a second object type label.

In this way, the system state data of all the target historical transactions can be input into the system state clustering model in a mode that one item of system state data of the target historical transactions is taken as one input quantity, and a second target type label which is output by the system state clustering model and represents the system state type of each target historical transaction is obtained.

For example, the number of target historical transactions may be 10 (i.e. system state data 1-10), and the above steps may refer to that the system state data 1 and the system state data 2 … … are respectively input into the system state clustering model by using the system state data 10 as an input amount, so as to obtain the second target type label corresponding to each of the system state data 1-10.

In this case, the determining, according to the training samples marked with the target type tag and the transaction state being the successful state, the proportion of the training samples marked with the target type tag in all training samples includes:

S132, determining a second transaction success rate of the transaction request according to the proportion of training samples marked with the second target type tag and the transaction state being the successful state in all the training samples marked with the second target type tag, wherein the target transaction success rate comprises the second transaction success rate.

For example, if the second target type tag output by the transaction state prediction model is B, and in the training samples of the system state prediction model, the samples of the tag B are 10000, and the samples of the transaction success are 1500, it may be determined that the transaction success rate corresponding to the first target type tag B is 15%, that is, the second transaction success rate corresponding to the transaction request is 15%.

In this case, whether to execute the transaction operation corresponding to the transaction request may be determined based on the magnitude relation between the second transaction success rate and the success rate threshold, and the specific policy is described with reference to the embodiment of step S14, which is not described in detail herein.

According to the technical scheme, the success rate of the transaction corresponding to the transaction request can be predicted based on the system state data of the target historical transaction, so that the occurrence frequency of the transaction failure phenomenon can be reduced.

In one possible implementation manner, the determining the second transaction success rate of the transaction request according to the proportion of training samples marked with the second target type tag and having the transaction state being a successful state in all training samples marked with the second target type tag (S132) includes:

for each of the second target type tags, determining a first duty cycle of training samples marked with such second target type tag in all training samples.

For example, the number of target historical transactions is 10, wherein the number of target historical transactions with the second target type tag of the system state data being a is 3, the number of target historical transactions with the second target type tag of the system state data being B is 4, and the number of target historical transactions with the second target type tag of the system state data being C is 3.

And when the success rate of the second transaction is determined, for the second target type label being A, the ratio of the sample with the second target type label being A in all training samples to the number of all training samples can be used as the first duty ratio. Similarly, a first duty cycle corresponding to the second object type tag B, the second object type tag C may also be determined.

Further, for each of the second target type tags, a second duty ratio of system state data belonging to such second target type tag is determined among all of the state data corresponding to the target history transactions.

Along with the above example, for the second target type tag being a, the ratio of the number of state data of which the second type tag of system state data in the target historical transaction is a to the total number of system state data in the target historical transaction may also be taken as the second duty cycle of the second target type tag a. Along the above example, the second duty cycle of the second object type tag a is 30%. Similarly, the second duty cycle of the second object type tag B may be determined to be 40% and the second duty cycle of the second object type tag C may be determined to be 30%.

Of course, in some embodiments, the second duty cycle F ₂ may also be determined by the following formula:

wherein F _2A is the second duty ratio of the second target type tag a, x is the total number of state data corresponding to the target historical transaction, and x _A is the number of system state data of the second target type tag a.

In this way, the weight corresponding to the system state data of each target historical transaction can be determined according to the first duty ratio and the second duty ratio of each target type tag. For example, the weight corresponding to each system state data may be calculated by the following formula:

Wherein F _t＝F_t1×F_t2,F_t1 is a first duty ratio of the second target type tag corresponding to the system state data t, and F _t2 is a second duty ratio of the second target type tag corresponding to the system state data t. a _t is the weight corresponding to the system state data t, NT is the time step, in this case the number of system state data input to the system state cluster model.

After determining the weight of the system state data for each target historical transaction, reference is made to the following equation:

And carrying out weighted summation on the weight corresponding to the system state data of each target historical transaction and the target proportion corresponding to the second target type label of the system state data to obtain the second transaction success rate. The target proportion is a training sample marked with the second target type label and the transaction state is a successful state, and the proportion of the training sample marked with the second target type label is occupied in all training samples. P _t is the second transaction success rate, P _w1 is a target proportion, and P _w1 may be different for different status data.

By adopting the technical scheme, the success rate of the transaction corresponding to the transaction request can be predicted based on the system state data of the target historical transaction, so that the occurrence frequency of the transaction failure phenomenon can be reduced.

In addition, in some implementation scenarios, a system state prediction model may also be established according to system state data of historical transactions, so that a success rate of the transactions may be predicted according to the system state data corresponding to the target historical transactions and the system state prediction model.

FIG. 4 is a flowchart of a blockchain network transaction method according to an exemplary embodiment of the present disclosure, where the method is based on FIG. 1, the state data includes system state data of the target historical transaction, the environment prediction model includes a system state prediction model, and the inputting the state data into the environment prediction model obtains a target type tag of the state data output by the environment prediction model, and includes:

S123, taking the system state data of all the target historical transactions as one input quantity of the system state prediction model, inputting the system state prediction model, and obtaining a third target type label which is output by the system state prediction model and represents the system state type.

For the system state data and the LSTM construction, please refer to fig. 2 and fig. 3 for description of the related content embodiments, which are not limited in this disclosure for brevity of description.

With respect to the training samples, the training samples of the system state prediction model are in one-to-one correspondence with historical transaction time periods of the blockchain network, and system state data corresponding to all historical transactions of the same blockchain node in one historical time period is used as one training sample.

After constructing the system state prediction model based on the LSTM, the determined training sample data can be input into the LSTM to obtain the prediction result of the LSTM. For example, for system state data 1-11, if the training samples input to the system state prediction model are system state data 1-10, the type tag of system state data 11 may be used as the type tag of the training samples. In this way, the correctness of the prediction result of the system state prediction model can be judged by comparing the prediction label output by the system state model with the type label of the training sample, and further relevant network parameters of the LSTM are updated until the model converges, so that training is completed.

s133, determining a third transaction success rate of the transaction request according to the proportion of training samples marked with the third target type tag and the transaction state being the successful state in all the training samples marked with the third target type tag. Wherein the target transaction success rate includes the third transaction success rate.

In this case, whether to execute the transaction operation corresponding to the transaction request may be determined based on the magnitude relation between the third transaction success rate and the success rate threshold, and the specific policy is described with reference to the embodiment of step S14, which is not described in detail herein.

According to the technical scheme, the transaction result can be predicted based on the system state data and the system state prediction model constructed based on the system state data, so that the occurrence rate of transaction failure phenomenon can be reduced.

FIG. 5 is a flowchart of a blockchain network transaction method according to an exemplary embodiment of the present disclosure, where the method is based on FIG. 1, the status data includes transaction status data of the target historical transaction, the environmental prediction model includes a transaction status cluster model, and the inputting the status data into the environmental prediction model obtains a target type tag of the status data output by the environmental prediction model, and includes:

S124, inputting the transaction state data of all the target historical transactions into the transaction state clustering model in a mode that the transaction state data of one target historical transaction is taken as an input quantity, and obtaining a fourth target type label which is output by the transaction state clustering model and represents the transaction state type of each target historical transaction.

The training samples of the transaction state clustering model are transaction state data corresponding to each historical transaction of the blockchain network, the transaction state clustering model is clustered into a plurality of class clusters aiming at all training samples in the training process, and the fourth target type label is a target class cluster to which the transaction state data representing the target historical transaction belongs.

For the transaction status data and the clustering method, please refer to the embodiments of fig. 2 and fig. 3 for description, and the disclosure is not repeated here.

S134, determining a fourth transaction success rate of the transaction request according to the proportion of training samples marked with the fourth target type tag and the transaction state being the success state in all the training samples marked with the fourth target type tag, wherein the target transaction success rate comprises the fourth transaction success rate.

In this case, whether to execute the transaction operation corresponding to the transaction request may be determined based on the magnitude relation between the fourth transaction success rate and the success rate threshold, and the specific policy is described with reference to the embodiment of step S14, which is not described in detail herein.

According to the technical scheme, the transaction result can be predicted based on the transaction state data and the transaction state prediction model constructed based on the transaction state data, so that the occurrence rate of the transaction failure phenomenon can be reduced.

It should be noted that, in the above embodiment, the success rate of the transaction is predicted by the transaction status data and the system status data, respectively, but those skilled in the art should appreciate that the methods shown in fig. 2 and 3 may be used in combination when implementing the present embodiment. For example, the first transaction success rate P _w2 may be obtained by the transaction state data and the transaction state prediction model, and the second transaction success rate P _w1 may be obtained by the system state data and the system state clustering model. In this way, the final transaction success rate can be obtained as the target transaction success rate by respectively setting weights for the first transaction success rate and the second transaction success rate. For example, the target transaction success rate P may be:

P＝αP_w2+(1-α)P_w1

Wherein α may be set according to application requirements, which is not limited by the present disclosure. Similarly, the method shown in fig. 4 and fig. 5 may be combined to synthesize the third transaction success rate and the fourth transaction success rate to predict the success rate of the transaction to be performed.

Furthermore, with respect to the training samples in the above embodiments, in some possible scenarios, the various models described in the above embodiments may also be continuously trained based on the input data. For example, when the state data is input to the model this time, the state data input this time may be regarded as the training sample in the above embodiment when the state data is input next time and the success rate of the transaction is calculated, which is not limited in the present disclosure.

In a possible implementation manner, the success rate of the transaction can be comprehensively predicted in combination with the methods shown in fig. 2 to 5, in which case, the state data includes transaction state data of the target historical transaction and system state data, the system state data is used for characterizing a system state at the occurrence time of the target historical transaction, and the environment prediction model includes a transaction state prediction model, a system state clustering model, a transaction state clustering model and a system state prediction model; accordingly:

the target transaction success rate comprises:

A first transaction success rate obtained based on the transaction state data and the transaction state prediction model; a second transaction success rate obtained based on the system state data and the system state clustering model; a third transaction success rate obtained based on the system state data and the system state prediction model; a fourth transaction success rate obtained based on the transaction state data and the transaction state clustering model; the determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold value comprises the following steps:

Regarding the various models and the calculation manners of the transaction success rates in the present embodiment, please refer to the embodiments illustrated in fig. 2 to 5, and the disclosure will not be repeated for brevity of description.

In addition, in some implementation scenarios, corresponding threshold data may be set for the first transaction success rate to the fourth transaction success rate, respectively, and the policy of the transaction may be determined by comparing each transaction success rate with the corresponding threshold data. For example, conduct transactions, cancel transactions, etc.

According to the embodiment, the transaction success rate is predicted through the transaction state data, the system state data and the four models, so that the accuracy of a prediction result is improved.

In addition, it should be noted that, regarding the calculation methods of the first transaction success rate, the third transaction success rate, and the fourth transaction success rate in the above embodiments, referring to the above description of the embodiment of step S132, in some embodiments, the calculation of the first transaction success rate, the third transaction success rate, and the fourth transaction success rate may also be performed based on the same inventive concept, which is not described herein.

FIG. 6 is a flowchart of a method for determining a cluster number of a cluster model according to an exemplary embodiment of the present disclosure, where the system state cluster model or the transaction state cluster model target cluster number is obtained as shown in FIG. 6, by:

In S61, a plurality of candidate cluster numbers are set.

For example, a range of candidate clusters [ K, K ] may be determined, resulting in a number of loops K-K.

In S62, for each candidate cluster number, clustering the data to be clustered according to the candidate cluster number until the cluster center converges.

For example, for a cluster sample set d= { x1, x2, x3,..once, xi,..once, xn }, comprising n transaction system state data, the number of cluster centers may be defined as k, so that k samples may be randomly selected as cluster centers, which may be (μ ₁,μ₂,μ₃,...,μ_k), for example.

After the cluster centers are selected, the distances from each cluster sample to each cluster center may be calculated, for example: d _ij＝||x_i-μ_j||². Where d _ij is the distance of sample x _i from cluster center μ _j. Thus, for cluster center μ _j, j=1, 2,3,..k, the cluster center can be recalculated:

In this way, through iterating the updating process of the cluster center, a converged cluster center can be finally obtained, and a cluster division result C= { C ₁,C₂,...,C_k } is output according to the converged cluster center.

In S63, inter-class distance values of a plurality of cluster categories corresponding to each candidate cluster number and intra-class distance values of respective data in each of the cluster categories are calculated.

Along the above example, the intra-class spacing may be:

Wherein d is an inter-class distance value corresponding to the cluster number k, which is equal to the maximum value of the sum of the average distance from the ith sample of the jth class to all samples in the class and the distance between the ith sample and the cluster center of the jth class cluster. For the mth sample of the j-th class,Mu _j is the cluster center of the j-th class cluster, n _j is the number of samples of the j-th class cluster, j=1, 2,3,..k.

In addition, the inter-class spacing may be:

Wherein D is an inter-class distance value corresponding to the cluster number k, which is equal to the minimum value of the sum of the average distance from the ith sample of the jth class to the samples in other classes and the cluster center distance. For the mth sample of the kth class,Mu _k is the cluster center of the kth class cluster, n _k is the number of samples of the kth class cluster, j=1, 2,3,..k.

In S64, the candidate cluster number corresponding to the maximum value of the differences between the inter-class distance values and the intra-class distance values is used as the target cluster number.

For example, for the candidate cluster number k, a difference d ^k between the inter-class distance value corresponding to the class candidate cluster number and the intra-class distance value corresponding to the class candidate cluster number may be calculated:

d^k＝D-d

Thus, the difference list d= [ d ^k,d^k+1,...,d^K ] can be obtained for the candidate cluster numbers K, k+1, …, K, so that the candidate cluster number corresponding to the maximum value of the differences between the inter-class pitch values and the intra-class pitch values can be used as the target cluster number. Of course, in some embodiments, the candidate cluster number corresponding to the minimum value in the difference between the intra-class distance value and the inter-class distance value may be used as the target cluster number, which is not limited in the disclosure. After determining the target cluster number, the data to be clustered (such as system state data and transaction state data) can be clustered according to the target cluster number, so as to obtain type labels corresponding to the data to be clustered.

According to the technical scheme, the target cluster number with better clustering effect can be determined from the candidate cluster numbers by setting the candidate cluster numbers and calculating the inter-class distance and the intra-class distance corresponding to the cluster numbers. Therefore, the target clustering number is adopted to cluster the to-be-clustered, and the accuracy of the clustering result can be improved.

Based on the same inventive concept, the present disclosure further provides a blockchain network transaction device, referring to a block diagram of the blockchain network transaction device shown in fig. 7, the device 700 includes:

A first obtaining module 701, configured to obtain, in response to a target node obtaining a transaction request, state data corresponding to a target historical transaction of the target node, where the target historical transaction is a historical transaction generated by the target node within a preset time period before a current time, or the target historical transaction is a preset number of historical transactions that are generated by the target node last before the target node obtains the transaction request;

A first input module 702, configured to input the state data into an environmental prediction model, to obtain a target type tag of the state data output by the environmental prediction model, where training samples of the environmental prediction model are obtained based on state data of historical transactions generated in a blockchain network, and each training sample is labeled with a type tag;

A first determining module 703, configured to determine, according to training samples marked with the target type tag and having a transaction status that is a successful status, a target transaction success rate of the transaction request according to a proportion of the training samples marked with the target type tag;

And a second determining module 704, configured to determine whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold.

the first input module 702 includes:

The first input sub-module is used for taking the transaction state data of all the target historical transactions as one input quantity of the transaction state prediction model, inputting the transaction state prediction model to obtain a first target type label which is output by the transaction state prediction model and is used for representing the transaction state type, wherein training samples of the transaction state prediction model are in one-to-one correspondence with historical transaction time periods of the blockchain network, and the transaction state data corresponding to all the historical transactions of the same blockchain node in one historical time period are used as one training sample;

the first determining module 703 includes:

The first determining submodule is used for determining a first transaction success rate of the transaction request according to the proportion of training samples marked with the first target type tag and the transaction state being the success state in all the training samples marked with the first target type tag, wherein the target transaction success rate comprises the first transaction success rate.

the first input module 702 includes:

the second input sub-module is used for inputting the system state clustering model in a mode that one item of system state data of the target historical transaction is taken as one input quantity, so as to obtain a second target type label which is output by the system state clustering model and represents the system state type of each target historical transaction, wherein a training sample of the system state clustering model is the system state data corresponding to each historical transaction of the blockchain network, the system state clustering model clusters into a plurality of class clusters aiming at all training samples in the training process, and the second target type label is a target class cluster to which the system state data representing the target historical transaction belongs;

the first determining module 703 includes:

And the second determining submodule is used for determining a second transaction success rate of the transaction request according to the proportion of training samples marked with the second target type tag and the transaction state being the successful state in all the training samples marked with the second target type tag, wherein the target transaction success rate comprises the second transaction success rate.

Optionally, the second determining sub-module includes:

a first determining subunit, configured to determine, for each of the second target type tags, a first duty ratio of training samples marked with such second target type tag in all training samples;

A second determining subunit, configured to determine, for each of the second target type tags, a second duty ratio of system state data belonging to the second target type tag, from among state data corresponding to all the target historical transactions;

a third determining subunit, configured to determine, according to the first duty ratio and the second duty ratio of each of the second target type tags, a weight corresponding to the system state data of each of the target historical transactions;

The calculating subunit is used for carrying out weighted summation on the weight corresponding to the system state data of each target historical transaction and the target proportion corresponding to the second target type label of the system state data to obtain the second transaction success rate;

the first input module 702 includes:

The third input sub-module is used for taking the system state data of all the target historical transactions as one input quantity of the system state prediction model, inputting the system state prediction model to obtain a third target type label which is output by the system state prediction model and is used for representing the system state type, wherein training samples of the system state prediction model are in one-to-one correspondence with the historical transaction time periods of the blockchain network, and the system state data corresponding to all the historical transactions of the same blockchain node in one historical time period is used as one training sample;

the first determining module 703 includes:

And the third determining submodule is used for determining a third transaction success rate of the transaction request according to the proportion of training samples marked with the third target type tag and the transaction state being the success state in all the training samples marked with the third target type tag, wherein the target transaction success rate comprises the third transaction success rate.

the first input module 702 includes:

A fourth input sub-module, configured to input transaction state data of all the target historical transactions into the transaction state clustering model in a manner that transaction state data of a target historical transaction is used as an input quantity, so as to obtain a fourth target type tag that is output by the transaction state clustering model and represents a transaction state type of each target historical transaction, a training sample of the transaction state clustering model is transaction state data corresponding to each historical transaction of the blockchain network, the transaction state clustering model clusters into a plurality of class clusters for all training samples in a training process, and the fourth target type tag is a target class cluster to which the transaction state data representing the target historical transaction belongs;

the first determining module 703 includes:

And the fourth determining submodule is used for determining a fourth transaction success rate of the transaction request according to the proportion of training samples marked with the fourth target type tag and the transaction state being the success state in all the training samples marked with the fourth target type tag, wherein the target transaction success rate comprises the fourth transaction success rate.

the target transaction success rate comprises:

The second determining module includes:

the first calculation sub-module is used for carrying out weighted summation on the first transaction success rate, the second transaction success rate, the third transaction success rate and the third transaction success rate to obtain a sum value;

The first execution sub-module is used for executing the transaction operation corresponding to the transaction request under the condition that the sum value is larger than the success rate threshold value;

And the second execution sub-module is used for revoking the transaction request under the condition that the sum value is smaller than or equal to the success rate threshold value.

Optionally, the apparatus further comprises: an execution module for determining a target cluster number of the system state cluster model or the transaction state cluster model, the execution module comprising:

the setting submodule is used for setting a plurality of candidate cluster numbers;

The clustering sub-module is used for clustering the data to be clustered according to the candidate cluster numbers aiming at each candidate cluster number until the cluster center converges;

a second calculation sub-module for calculating inter-class distance values of a plurality of cluster categories corresponding to each candidate cluster number and intra-class distance values of each data in each cluster category;

And the third execution sub-module is used for taking the candidate cluster number corresponding to the maximum value in the difference value of the inter-class distance value and the intra-class distance value as the target cluster number.

The specific manner in which the various modules perform the operations in the apparatus of the above embodiments have been described in detail in connection with the embodiments of the method, and will not be described in detail herein.

The present disclosure also provides a computer readable storage medium having stored thereon a computer program which when executed by a processor performs the steps of the blockchain network transaction method provided by the present disclosure.

The present disclosure also provides an electronic device, including:

A memory having a computer program stored thereon;

a processor for executing the computer program in the memory to implement the steps of the blockchain network transaction method provided by the present disclosure.

Fig. 8 is a block diagram of an electronic device 800, which electronic device 800 may, for example, act as a blockchain node, shown in accordance with an exemplary embodiment. As shown in fig. 8, the electronic device 800 may include: a processor 801, a memory 802. The electronic device 800 may also include one or more of a multimedia component 803, an input/output (I/O) interface 804, and a communication component 805.

The processor 801 is configured to control overall operation of the electronic device 800 to perform all or part of the steps of the blockchain network transaction method described above. The memory 802 is used to store various types of data to support operation at the electronic device 800, which may include, for example, instructions for any application or method operating on the electronic device 800, as well as application-related data, such as tile data, certificates, pictures, video, and the like. The Memory 802 may be implemented by any type or combination of volatile or non-volatile Memory devices, such as static random access Memory (Static Random Access Memory, SRAM for short), electrically erasable programmable Read-Only Memory (ELECTRICALLY ERASABLE PROGRAMMABLE READ-Only Memory, EEPROM for short), erasable programmable Read-Only Memory (Erasable Programmable Read-Only Memory, EPROM for short), programmable Read-Only Memory (Programmable Read-Only Memory, PROM for short), read-Only Memory (ROM for short), magnetic Memory, flash Memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. Wherein the screen may be, for example, a touch screen, the audio component being for outputting and/or inputting audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted through the communication component 805. The audio component may be at least one speaker for outputting audio signals. The I/O interface 804 provides an interface between the processor 801 and other interface modules, which may be a keyboard, mouse, buttons, etc. These buttons may be virtual buttons or physical buttons. The communication component 805 is used for wired or wireless communication between the electronic device 800 and other devices. Wireless Communication, such as Wi-Fi, bluetooth, near Field Communication (NFC) for short, 2G, 3G, 4G, NB-IOT, eMTC, or other 5G, etc., or one or a combination of more of them, is not limited herein. The corresponding communication component 805 may thus comprise: wi-Fi module, bluetooth module, NFC module, etc.

In an exemplary embodiment, the electronic device 800 may be implemented by one or more Application-specific integrated circuits (ASICs), digital Signal Processors (DSPs), digital signal processing devices (DIGITAL SIGNAL Processing Device DSPDs), programmable logic devices (Programmable Logic Device PLDs), field programmable gate arrays (Field Programmable GATE ARRAY FPGAs), controllers, microcontrollers, microprocessors, or other electronic components for performing the blockchain network transaction methods described above.

In another exemplary embodiment, a computer readable storage medium is also provided that includes program instructions that when executed by a processor implement the steps of the blockchain network transaction method described above. For example, the computer readable storage medium may be the memory 802 described above including program instructions executable by the processor 801 of the electronic device 800 to perform the blockchain network transaction method described above.

In another exemplary embodiment, a computer program product is also provided, the computer program product comprising a computer program executable by a programmable apparatus, the computer program having code portions for performing the blockchain network transaction method described above when executed by the programmable apparatus.

The preferred embodiments of the present disclosure have been described in detail above with reference to the accompanying drawings, but the present disclosure is not limited to the specific details of the embodiments described above, and various simple modifications may be made to the technical solutions of the present disclosure within the scope of the technical concept of the present disclosure, and all the simple modifications belong to the protection scope of the present disclosure.

In addition, the specific features described in the foregoing embodiments may be combined in any suitable manner, and in order to avoid unnecessary repetition, the present disclosure does not further describe various possible combinations.

Moreover, any combination between the various embodiments of the present disclosure is possible as long as it does not depart from the spirit of the present disclosure, which should also be construed as the disclosure of the present disclosure.

Claims

1. A blockchain network transaction method, comprising:

determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold;

wherein the status data includes one or more of transaction status data of the target historical transaction, system status data of the target historical transaction;

the transaction state data comprises one or more of transaction concurrency number, transaction response time, transaction throughput, transaction timeout setting time, block-out time setting information, block size setting information, last transaction result identification and current transaction result identification;

the system state data comprises one or more of CPU occupancy rate, network occupancy bandwidth, memory occupancy rate, disk usage rate, CPU occupancy change rate, network occupancy bandwidth change rate, memory occupancy change rate and disk usage change rate of a central processing unit.

2. The method of claim 1, wherein the status data comprises transaction status data of the target historical transaction, and the environmental prediction model comprises a transaction status prediction model, and accordingly:

3. The method of claim 1, wherein the state data comprises system state data characterizing a system state at a time of occurrence of the target historical transaction, the environmental prediction model comprising a system state cluster model, and accordingly:

4. The method of claim 3, wherein determining the second transaction success rate of the transaction request based on the proportion of training samples marked with the second target type tag and the transaction status being a successful status, comprises:

determining the weight corresponding to the system state data of each target historical transaction according to the first duty ratio and the second duty ratio of each second target type tag;

The weight corresponding to the system state data of each target historical transaction and the target proportion corresponding to the second target type label of the system state data are weighted and summed to obtain the second transaction success rate;

5. The method of claim 1, wherein the status data comprises system status data of the target historical transaction, and the environmental prediction model comprises a system status prediction model, and accordingly:

6. The method of claim 1, wherein the status data comprises transaction status data of the target historical transaction, and the environmental prediction model comprises a transaction status cluster model, corresponding to:

7. The method of claim 1, wherein the status data comprises transaction status data of the target historical transaction and system status data characterizing a system status at a time of occurrence of the target historical transaction, the environmental prediction model comprising a transaction status prediction model, a system status cluster model, a transaction status cluster model, and a system status prediction model; accordingly:

the target transaction success rate comprises:

8. The method according to claim 3 or 6, wherein the target cluster number of the system state cluster model or the transaction state cluster model is obtained by:

setting a plurality of candidate cluster numbers;

9. The method according to any of claims 1 to 7, wherein the transaction request is a cross-chain transaction request, and accordingly the target historical transaction is a historical cross-chain transaction generated by the target node within a preset time period before the current moment, or the target historical transaction is a preset number of historical cross-chain transactions generated by the target node most recently before the transaction request is acquired, and training samples of the environment prediction model are obtained based on state data of the historical cross-chain transactions generated in a blockchain network.

10. A blockchain network transaction device, comprising:

the second determining module is used for determining whether to execute the transaction operation corresponding to the transaction request according to the magnitude relation between the target transaction success rate and the success rate threshold;

11. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method according to any one of claims 1-9.

12. An electronic device, comprising:

A memory having a computer program stored thereon;

A processor for executing the computer program in the memory to carry out the steps of the method of any one of claims 1-9.