CN113326879A

CN113326879A - Service data monitoring method and device

Info

Publication number: CN113326879A
Application number: CN202110597265.4A
Authority: CN
Inventors: 伍日杰; 王志远
Original assignee: WeBank Co Ltd
Current assignee: WeBank Co Ltd
Priority date: 2021-05-31
Filing date: 2021-05-31
Publication date: 2021-08-31
Anticipated expiration: 2041-05-31
Also published as: CN113326879B; WO2022252573A1

Abstract

The invention discloses a method and a device for monitoring service data, wherein the method comprises the following steps: the method comprises the steps of obtaining first business data in a preset time period and second business data in a historical time period related to the preset time period, wherein the first business data and the second business data aim at the same business index to generate images to be identified representing the first business data and the second business data, inputting the images to be identified into a convolutional neural network model, and determining a monitoring result of the business data in the preset time period under the business index, wherein the convolutional neural network model is obtained by training according to a historical identification image with a historical monitoring result label, and the historical monitoring result is determined according to the relation between the first business data and the second business data in the historical identification image. The convolutional neural network model integrates the condition of the service data of each historical time period, and can be suitable for monitoring the service data at all time periods, so that the dynamic monitoring of the service data is realized, and the accuracy of the service data monitoring is improved.

Description

Service data monitoring method and device

Technical Field

The invention relates to the field of financial technology (Fintech), in particular to a method and a device for monitoring business data.

Background

With the development of computer technology, more and more technologies (such as block chains, cloud computing or big data) are applied in the financial field, the traditional financial industry is gradually changing to the financial technology, and big data technology is no exception, but higher requirements are also put forward on big data technology due to the security and real-time requirements of the financial and payment industries.

Currently, a method for monitoring service data generally includes comparing service data in a current period with service data in a plurality of historical periods, for example, taking a day as a period unit, if monitoring whether the service data in the current period (5 months and 8 days) is abnormal, comparing the service data in the current period with service data in a plurality of historical periods (e.g., 5 months and 5 days, 5 months and 6 days, and 5 months and 7 days) before the current period, and if fluctuation (increase or decrease) of the service data in the current period is within a threshold range, determining that the service data in the current period is normal data.

However, the threshold range needs manual setting and adjustment, the threshold range for a cycle cannot be adapted to the whole period of the cycle, the service data cannot be dynamically monitored, and the accuracy is not high. For example, in a day cycle unit, in a service under a certain scene, in each cycle, more service data is 8:00-17:00, less service data is 17:00-8:00, and only in a time period of 17:00-18:00, the service data is rapidly reduced and exceeds a threshold range, and at this time, the generated exception is wrong and is actually normal service data.

Therefore, a method for monitoring service data is needed to dynamically monitor service data and improve the accuracy of monitoring service data.

Disclosure of Invention

The embodiment of the invention provides a method and a device for monitoring service data, which are used for dynamically monitoring the service data and improving the accuracy of monitoring the service data.

In a first aspect, an embodiment of the present invention provides a method for monitoring service data, including:

acquiring first service data in a preset time period and second service data in a historical time period associated with the preset time period; the first service data and the second service data aim at the same service index;

generating images to be identified representing the first business data and the second business data;

inputting the image to be recognized into a convolutional neural network model, and determining a monitoring result of the service data in the preset time period under the service index; the convolutional neural network model is obtained by training according to a historical recognition image with a historical monitoring result label; the historical monitoring result is determined according to the relation between the first service data and the second service data in the historical identification image.

The monitoring threshold value in the prior art is manually set according to experience; the convolutional neural network model is obtained by training according to historical recognition images; each historical identification image comprises first business data (current business data) and second business data (historical business data) and is provided with a label for representing a historical monitoring result; training the first convolution neural network model through the historical recognition image, so that the first convolution neural network model can recognize the relation between the first business data and the second business data and determine whether the first business data is abnormal or not according to the relation between the first business data and the second business data. Each historical identification image in the application represents various historical time periods, so that the convolutional neural network model is actually the condition of fusing the service data of each historical time period, and can be suitable for monitoring the service data in the whole time period. Therefore, the dynamic monitoring of the service data can be realized, and the accuracy of the service data monitoring is improved.

Optionally, the service indicator is at least one of the following: the service transaction amount, the average service consumption time and the service success rate;

the historical period comprises a ring ratio period of the preset period and/or a parity period of the preset period.

In the technical scheme, the service indexes are classified into different types, the comprehensiveness of service data monitoring is improved, and the ring ratio time period and the same ratio time period of the preset time period are more relevant to the preset time period, so that the service data can be compared better, and the accuracy of service data monitoring is improved.

Optionally, generating an image to be identified that represents the first service data and the second service data includes:

generating a coordinate system by taking time as an abscissa and a service index as an ordinate;

determining a first curve of the first service data in the coordinate system and a second curve of the second service data in the coordinate system to obtain a curve graph;

and preprocessing the graph to determine the image to be identified.

In the prior art, whether the service data is abnormal is determined according to the difference between the current service data and the historical data, and if the difference is not verified, whether the difference is abnormal is determined; in the application, a first difference between first business data and second business data is visually expressed in an image form, whether the first difference is abnormal or not is determined according to a second difference of a historical recognition image in a convolutional neural network model, whether the first business data is abnormal or not is determined according to whether the first difference is abnormal or not, and accordingly, twice monitoring is achieved, and accuracy of business data monitoring is improved.

Optionally, preprocessing the graph to determine the image to be recognized, including:

generating a picture by the graph;

the resolution of the picture is zoomed into a preset resolution, and the pixel value of any pixel point in the image to be identified is determined according to the following formula (1) to obtain the image to be identified;

wherein f (P) is the pixel value of any pixel point P in the image to be identified, (x, y) is the coordinate value of the pixel point P, and (x1, y1) is the coordinate value of an adjacent pixel point Q11 positioned at the lower left corner of the pixel point P in the picture; (x1, y2) is the coordinate value of the adjacent pixel point Q12 which is positioned at the upper left corner of the pixel point P in the picture; (x2, y1) is the coordinate value of the adjacent pixel point located in the lower right corner Q21 of the pixel point P in the picture; (x2, y2) is the coordinate value of the adjacent pixel point Q22 which is positioned at the upper right corner of the pixel point P in the picture; f (Q11) is the pixel value of the pixel Q11, f (Q12) is the pixel value of the pixel Q12, f (Q21) is the pixel value of the pixel Q21, and f (Q22) is the pixel value of the pixel Q22.

In the technical scheme, due to different values of the service data, the peak values in the coordinate system have different sizes, so that the generated pictures have different resolutions, and the convolutional neural network model has abnormal monitoring conditions, so that the resolution of the images to be identified is unified through preprocessing, and the accuracy of monitoring the service data is improved.

Optionally, after the image to be recognized is input into a convolutional neural network model and a monitoring result of the service data in the preset time period under the service index is determined, the method further includes:

and determining a comprehensive monitoring result of the service data in the preset time period according to the preset weight of each service index and the monitoring result of the service data in the preset time period under each service index.

In the technical scheme, the detection results obtained by each service index are aggregated according to the preset weight to obtain the comprehensive monitoring result, so that the abnormal service data of a certain service index is prevented, and the first service data is judged to be abnormal by mistake under the condition that the service data of other service indexes are normal, so that the accuracy of monitoring the service data is improved.

Optionally, the convolutional neural network model is used for performing N classification;

inputting the image to be recognized into a convolutional neural network model, and determining a monitoring result of the service data in the preset time period, wherein the monitoring result comprises the following steps:

inputting the image to be identified into a convolutional neural network model to obtain a classification result;

and determining a monitoring result corresponding to the classification result according to the comparison relation between the N classification and the M classification monitoring result, wherein N is larger than M.

In the above technical solution, the comparison relationship between the N-class and M-class monitoring results is used to determine the classification result of the image to be identified, so as to determine whether the image to be identified is abnormal, wherein the comparison relationship between the N-class and M-class monitoring results can be preset by a user according to experience, so as to increase flexibility of monitoring the service data.

Optionally, when it is determined that the number of the monitoring results of the error anomalies is greater than the number threshold, the convolutional neural network model is trained according to the image to be recognized and the correction label corresponding to each error anomaly, so as to obtain an updated convolutional neural network model.

In the above technical scheme, for the abnormal service data determined by the convolutional neural network model, if the user determines that the abnormal service data is normal, that is, the convolutional neural network model is judged incorrectly, the abnormal service data is recorded, and when the number of the monitoring results of the incorrect abnormality is greater than a number threshold, the convolutional neural network model is optimally trained according to the incorrectly abnormal service data, so that the identification accuracy of the convolutional neural network model is increased, and the accuracy of monitoring the service data is increased.

In a second aspect, an embodiment of the present invention provides a device for monitoring service data, including:

the acquisition module is used for acquiring first service data in a preset time period and second service data of a historical time period associated with the preset time period; the first service data and the second service data aim at the same service index;

the processing module is used for generating images to be identified representing the first service data and the second service data;

Optionally, the processing module is specifically configured to:

and preprocessing the graph to determine the image to be identified.

Optionally, the processing module is specifically configured to:

generating a picture by the graph;

Optionally, the processing module is further configured to:

the processing module is specifically configured to:

Optionally, the processing module is further configured to:

and when the number of the monitoring results of the error abnormity is determined to be larger than the number threshold, training the convolutional neural network model according to the image to be recognized and the correction label corresponding to each error abnormity to obtain an updated convolutional neural network model.

In a third aspect, an embodiment of the present invention further provides a computer device, including:

a memory for storing program instructions;

and the processor is used for calling the program instruction stored in the memory and executing the monitoring method of the business data according to the obtained program.

In a fourth aspect, an embodiment of the present invention further provides a computer-readable storage medium, where the computer-readable storage medium stores computer-executable instructions, and the computer-executable instructions are configured to enable a computer to execute the foregoing service data monitoring method.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the description of the embodiments will be briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on these drawings without creative efforts.

FIG. 1 is a system architecture diagram according to an embodiment of the present invention;

fig. 2 is a schematic flow chart of a method for monitoring service data according to an embodiment of the present invention;

FIG. 3 is a schematic diagram of a graph according to an embodiment of the present invention;

FIG. 4 is a diagram illustrating a resolution scaling calculation according to an embodiment of the present invention;

fig. 5 is a schematic diagram of an image to be recognized according to an embodiment of the present invention;

FIG. 6 is a schematic diagram of a convolutional neural network model according to an embodiment of the present invention;

fig. 7 is a schematic structural diagram of a service data monitoring device according to an embodiment of the present invention.

Detailed Description

In order to make the objects, technical solutions and advantages of the present invention clearer, the present invention will be described in further detail with reference to the accompanying drawings, and it is apparent that the described embodiments are only a part of the embodiments of the present invention, not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

In the prior art, methods for monitoring service data are generally divided into the following two types:

first, the traffic data of the current period is compared with the traffic data of a plurality of historical periods. Taking a service index as a service transaction amount for example, taking a day as a cycle unit, monitoring service data of a current cycle (5 month and 8 days), comparing the service data of a historical cycle of the current cycle (including a cycle history, such as 5 month and 7 days, and a month comparison history, such as 4 month and 8 days), setting a cycle threshold as an upper limit and a lower limit of 30% (that is, the service transaction amount of the current cycle is not more than 30% of the service transaction amount compared with the service transaction amount of the cycle history), setting an upper limit of a comparison threshold as 20% (that is, the service transaction amount of the current cycle is not more than 20% of the service transaction amount compared with the service transaction amount of the month comparison history, and the service transaction amount is not more than 40%) and setting a lower limit of a comparison threshold as 40% (that is, the service transaction amount of the current cycle is not more than 40% of the service transaction amount compared with the service transaction amount of the month comparison history, if the threshold is exceeded, determining the service data of the current period as abnormal data.

Secondly, the service data range is determined by fitting a curve. Taking the service index as the service transaction amount as an example, for example, a fitting curve is generated according to the service data in the history period, the threshold range of the service transaction amount in the current period is determined to be 80-100 according to the fitting curve, and if the service transaction amount in the current period is not in the threshold range, the service data in the current period is determined to be abnormal data.

However, in the first method, the threshold needs to be manually set and adjusted, the threshold range for a period cannot be adapted to the whole period of the period, the service data cannot be dynamically monitored, the accuracy is not high, and a large number of false exceptions may occur when the traffic volume is extremely small, for example, the historical traffic volume is 1, the current traffic volume is 0, the current traffic volume is decreased by 100%, and the current traffic volume exceeds the threshold, and it is determined that the current traffic data is abnormal.

In the second method, the fitted curve cannot be dynamically monitored according to the actual conditions of the service data, for example, the fluctuation of the historical service data is large, the accuracy of the threshold range determined by the fitted curve is extremely small, and the threshold range cannot be optimized according to the determined error abnormal service data, so that the threshold range cannot be dynamically determined.

Fig. 1 illustrates an exemplary system architecture to which embodiments of the present invention are applicable, which includes a server 100, where the server 100 may include a processor 110, a communication interface 120, and a memory 130.

The communication interface 120 is configured to obtain first service data in a preset time period and second service data in a historical time period associated with the preset time period.

The processor 110 is a control center of the server 100, connects various parts of the entire server 100 using various interfaces and routes, performs various functions of the server 100 and processes data by operating or executing software programs and/or modules stored in the memory 130 and calling data stored in the memory 130. Alternatively, processor 110 may include one or more processing units.

The memory 130 may be used to store software programs and modules, and the processor 110 executes various functional applications and data processing by operating the software programs and modules stored in the memory 130. The memory 130 may mainly include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function, and the like; the storage data area may store data created according to a business process, and the like. Further, the memory 130 may include high speed random access memory, and may also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device.

It should be noted that the structure shown in fig. 1 is only an example, and the embodiment of the present invention is not limited thereto.

Based on the above description, fig. 2 exemplarily illustrates a flow diagram of a method for monitoring business data according to an embodiment of the present invention, where the flow may be executed by a device for monitoring business data.

As shown in fig. 2, the process specifically includes:

step 210, acquiring first service data in a preset time period and second service data in a historical time period associated with the preset time period.

In the embodiment of the present invention, the first service data and the second service data are for the same service index, for example, the first service data and the second service data are both service transaction amounts.

Step 220, generating images to be identified representing the first service data and the second service data.

In the embodiment of the present invention, the basis for determining whether the first service data is abnormal data includes the first service data and the second service data, and specifically, the first difference between the first service data and the second service data is visually expressed in the form of an image.

Step 230, inputting the image to be recognized into a convolutional neural network model, and determining a monitoring result of the service data in the preset time period under the service index.

In the embodiment of the invention, the convolutional neural network model is obtained by training according to a historical recognition image with a historical monitoring result label, and the historical monitoring result is determined according to the relation between first service data and second service data in the historical recognition image.

In step 210, the service index includes multiple categories, specifically, the service index is at least one of the following: traffic volume, average traffic time and traffic success rate.

In an implementation manner, the service index may be determined according to a service log collected at the service interface, for example, by a collection module (e.g., agent), the service log is collected at a preset service interface, and then the service index is determined according to the collected service log, as shown in table 1 below.

TABLE 1

The status code is used to indicate whether the service data is successful (i.e. normal), for example, the status code 200 indicates that the service data is normal, and the status code 500 indicates that the service data is abnormal. The service success rate is determined according to the normal service quantity and the total service quantity in a preset time period, for example, the ratio of the normal service quantity to the total service quantity is determined as the service success rate.

In the embodiment of the invention, the historical time period comprises a ring ratio time period of a preset time period and/or a comparation time period of the preset time period. For example, the preset time period is 7 months and 7 days 12:00-12:10, the ring ratio time period is 7 months and 7 days 11:50-12:00, and the comparable time period is 6 months and 7 days 12:00-12: 10. Wherein, the year-to-year period can be divided into week-to-year, month-to-month, year-to-year, and the like, for example, the preset period is 12:00-12:10 at 7/2021, the week-to-year period is 12:00-12:10 at 6/30/2021, the month-to-year period is 12:00-12:10 at 6/7/2021, and the year-to-year period is 12:00-12:10 at 7/2020/7. Again, the specific classifications of parity are not limiting.

In step 220, a coordinate system is determined according to a preset time period, the first service data and the second service data, a graph representing the first service data and the second service data is determined in the coordinate system, and an image to be identified is obtained according to the graph.

Specifically, a coordinate system is generated by taking time as an abscissa and a service index as an ordinate, a first curve of first service data in the coordinate system and a second curve of second service data in the coordinate system are determined to obtain a curve graph, and then the curve graph is preprocessed to determine an image to be identified.

In the embodiment of the present invention, the minimum unit and the peak value of the abscissa in the coordinate system are determined according to a preset time period, for example, the preset time period is 10 hours, and the abscissa is in the minimum unit of hours, or the minimum unit of half an hour, and the peak value is 10. The minimum unit and the peak value of the ordinate in the coordinate system are determined according to the service index, for example, the ordinate is the service success rate, the ordinate takes the success rate of 10% as the minimum unit, or the success rate of 20% as the minimum unit, and the peak value is 100%. It should be noted that the value of the minimum unit is only an example, and is not specifically limited herein.

In order to better describe the above technical solution, the service index Wie is used to describe the service transaction amount in the following specific example.

Example 1

The preset time period is 10 minutes, and the first service data in the preset time period is acquired, as shown in table 2 below.

TABLE 2

According to the same method, second service data of a historical period associated with the preset period is obtained, wherein the second service data comprises ring ratio historical service data and week ratio historical service data of the preset period, as shown in the following tables 3 and 4.

TABLE 3

TABLE 4

Now, with minutes as the minimum time unit, the summarized traffic data can be obtained according to the data in tables 2 to 4, as shown in table 5 below.

TABLE 5

	1	2	3	4	5	6	7	8	9	10
											A	70	75	71	74	69	79	68	61	72	84
B	61	79	59	78	65	66	74	55	73	77
											C	74	83	72	82	73	75	77	64	74	72

Wherein, a represents the first service data, B represents the ring ratio historical service data, and C represents the week and year ratio historical service data, and then a coordinate system is generated according to table 5, since the maximum data in table 5 is 84, the peak value of the ordinate in the coordinate system can be 85, 90, etc., and the peak value in the embodiment of the invention is 90.

Then, a graph is determined according to a coordinate system, as shown in fig. 3, fig. 3 exemplarily shows a schematic diagram of a graph, wherein the graph comprises first traffic data, ring ratio historical traffic data and week ratio historical traffic data, t represents time and is a unit of minutes, and n represents transaction traffic.

It should be noted that, the distinguishing of the first service data, the ring ratio historical service data, and the week ratio historical service data in the graph may be distinguished according to a line format, such as a straight line, a dotted line, and the like, and may also be distinguished according to a color of the line, for example, the first service data is a red straight line, the ring ratio historical service data is a blue straight line, and the week ratio historical service data is a green straight line, which is not specifically limited herein.

In the embodiment of the invention, after the curve graph is obtained, the curve graph is used for generating the picture, and the image to be identified is obtained according to the picture.

Specifically, for the graph generated picture, the coordinate system in fig. 3 is deleted, the picture is generated only according to the graph, then the picture is zoomed, the resolution of the picture is zoomed to a preset resolution, and then the image to be identified is obtained.

And for the obtained image to be identified, aiming at any coordinate point of the image to be identified with preset resolution, determining a pixel value of the coordinate point according to the pixel values of the adjacent coordinate points of the coordinate point in the picture, and obtaining the image to be identified.

Generating a picture by the curve graph;

the method comprises the steps of scaling the resolution of a picture to be a preset resolution, and determining the pixel value of any pixel point in an image to be identified according to the following formula (1) to obtain the image to be identified;

wherein f (P) is the pixel value of any pixel point P in the image to be identified, (x, y) is the coordinate value of the pixel point P, and (x1, y1) is the coordinate value of an adjacent pixel point Q11 positioned at the lower left corner of the pixel point P in the picture; (x1, y2) is the coordinate value of the adjacent pixel point Q12 located at the upper left corner of the pixel point P in the picture; (x2, y1) is the coordinate value of the adjacent pixel point in the picture at the lower right corner Q21 of the pixel point P; (x2, y2) is the coordinate value of the adjacent pixel point Q22 located at the upper right corner of the pixel point P in the picture; f (Q11) is the pixel value of the pixel Q11, f (Q12) is the pixel value of the pixel Q12, f (Q21) is the pixel value of the pixel Q21, and f (Q22) is the pixel value of the pixel Q22.

For example, the preset resolution is 224 pixels by 224 pixels, and the peak values of the coordinate system may be different and the heights or widths of the graphs may be different due to different maximum values of the service data in the preset time period, so that the resolutions of the pictures generated according to the graphs may be different, and therefore, in order to improve the monitoring determination rate, the resolution of the pictures is preset to the preset resolution.

If the resolution of the picture is greater than the preset resolution, the picture resolution needs to be reduced to the preset resolution, and if the resolution of the picture is less than the preset resolution, the picture resolution needs to be enlarged to the preset resolution.

When the resolution scaling is carried out, all pixel points in the image to be recognized are sequentially processed, primary calculation is respectively carried out in two directions, namely the X-axis direction and the Y-axis direction, the pixel value of the pixel point to be solved is obtained through interpolation of four adjacent pixel points, and the weight of the coordinate pixel closer to the coordinate position of the pixel point to be solved is larger in the calculation process.

Fig. 4 shows an exemplary schematic diagram of a resolution scaling calculation in conjunction with the above equation (1), and the following fig. 4 illustrates an example of obtaining an image to be recognized.

Example 2

As shown in fig. 4, Q11, Q12, Q21 and Q22 are four adjacent pixels of the pixel P in fig. 4, the pixel P is a pixel to be solved, and the formula (1) can be divided into the following formula (2), formula (3) and formula (4), and the specific steps are as follows:

1. f (R1) is obtained from the following formula (2) by Q11(x1, y1), Q21(x2, y1), and f (R2) is obtained from the following formula (3) by Q12(x1, y2), Q22(x2, y 2);

wherein f (R1) is the pixel value of the pixel P on the abscissa for the pixel points Q11 and Q21, f (R2) is the pixel value of the pixel P on the abscissa for the pixel points Q12 and Q22, and if the decimal pixel value exists after calculation, the calculation result is rounded.

2. P was obtained from R1(x, y1), R2(x, y2) according to the following calculation formula (4).

Therefore, the zooming of the picture can be realized, and then the zoomed picture is determined as the image to be identified. It should be noted that, in an implementable manner, a pixel value of the pixel point P on the ordinate with respect to the pixel points Q11, Q21, Q12, and Q22 may be determined, and then the pixel value of the pixel point P on the ordinate may be determined according to the following formula (5) and formula (6), and then the pixel value of the pixel point P on the ordinate with respect to the pixel points Q11, Q21, Q12, and Q22 may be determined according to the following formula (7).

In another implementable manner, a pixel value of any pixel point in the image to be recognized may also be determined according to a ratio of horizontal and vertical coordinates of the picture resolution to a preset resolution, for example, the picture resolution is mxn, the preset resolution is axb, the side length ratios are m/a and n/b, for any pixel point (i, j) in the image to be recognized, the pixel point of the corresponding picture is (im/a, jn/b), if im/a, jn/b is a non-integer, the pixel point of the corresponding picture is determined according to a rounding manner, and the pixel value of the pixel point is used as the pixel value of the pixel point (i, j), so as to obtain the pixel values of all the pixel points in the image to be recognized, and then the image to be recognized is determined. Therefore, in the embodiment of the present invention, the method for scaling the picture is not specifically limited.

In step 230, the convolutional neural network model is trained from the historical recognition image with the historical monitoring result label, and fig. 5 exemplarily shows a schematic diagram of an image to be recognized, in conjunction with fig. 3 and the above formula (1). The historical recognition image is similar to that shown in fig. 5, and operation and maintenance personnel give a historical monitoring result label to the historical recognition image, so that the convolutional neural network model is trained, and supervised machine learning is realized.

In one implementable approach, the convolutional neural network model can be a VggNet convolutional neural network model, a google lenet convolutional neural network model, or the like.

In the embodiment of the invention, the convolutional neural network model is an AlexNet convolutional neural network model.

And further, the convolutional neural network model is used for carrying out N classification, inputting the image to be recognized into the convolutional neural network model to obtain a classification result, and determining a monitoring result corresponding to the classification result according to the comparison relation between the N classification and the M types of monitoring results, wherein N is larger than M.

In the embodiment of the present invention, fig. 6 exemplarily shows a schematic diagram of a convolutional neural network model, as shown in fig. 6, AlexNet includes 5 convolutional layers (conv), 3 fully connected layers (full connected), the model output is 1000 digital values corresponding to 1000 classes, and the output result is converted into a decimal-corresponding traffic state result multiple class probability between 0 and 1 through a softmax function.

For example, the comparison relationship between the N classification and the M-class monitoring result may be a value preset by the operation and maintenance personnel according to experience, for example, the N classification includes (M1, … …, M1000), and the detection result is classified into 5 types including normal service, slight abnormal service, normal abnormal service, major abnormal service, and major abnormal service. Wherein, (m1, m2, … …, m200) corresponds to normal service, (m201, m202, … …, m400) corresponds to slight abnormal service, (m401, m402, … …, m600) corresponds to ordinary abnormal service, (m601, m602, … …, m800) corresponds to major abnormal service, and (m801, m802, … …, m1000) corresponds to major abnormal service.

In an implementation manner, for different types of service indexes, a comprehensive monitoring result can be determined according to weights corresponding to the service indexes, and whether the first service data is abnormal or not is determined according to the comprehensive monitoring result.

Specifically, an image to be recognized is input into the convolutional neural network model, after a monitoring result of the service data in the preset time period under the service indexes is determined, a comprehensive monitoring result of the service data in the preset time period is determined according to the preset weight of each service index and the monitoring result of the service data in the preset time period under each service index.

For example, in combination with the above technical solutions, weights may also be preset for the types of the detection results, for example, the normal weight of the service is 0.9, the slight abnormal weight of the service is 0.8, the normal abnormal weight of the service is 0.6, the major abnormal weight of the service is 0.3, and the major abnormal weight of the service is 0.1.

The preset weight of the service index may be, the service success rate weight is 0.5, the service average time consumption weight is 0.3, and the service transaction amount weight is 0.2.

Taking an example in combination with example 1, now, for 3 service indexes (service transaction amount, service average consumed time, and service success rate) of the first service data in a preset time period, the monitoring results determined according to the convolutional neural network model are respectively normal service, slight service abnormality, and normal service, and then a comprehensive monitoring result can be obtained according to a preset weight, where z is (0.6 × 0.2) + (0.8 × 0.3) + (0.9 × 0.5) ═ 0.81, and if 0.81 is greater than an abnormal threshold, it is determined that the first service data is normal data.

It should be noted that the preset weight and the abnormal threshold are set by a person based on experience, and are not specifically limited herein.

Illustratively, for the abnormal service data determined by the convolutional neural network model, if the operation and maintenance personnel determine that the abnormal service data is abnormal, the convolutional neural network model is optimized.

Specifically, when the number of the monitoring results of the error anomalies is determined to be larger than the number threshold, the convolutional neural network model is trained according to the image to be recognized and the correction label corresponding to each error anomaly, and the updated convolutional neural network model is obtained.

For example, when the abnormal service data a1 determined by the convolutional neural network model is determined to be abnormal, the operation and maintenance personnel marks the service data a1 when determining that the abnormal service data a1 is abnormal, and when determining that the number of the abnormal service data (such as 1001, including a1, … …, a1001) is greater than 1000 (number threshold), the operation and maintenance personnel uses the service data a1, … …, a1001 as training samples to perform optimization training on the convolutional neural network model to increase the recognition confidence of the convolutional neural network model and reduce the probability of determining the abnormal service data by the convolutional neural network model.

In the embodiment of the present invention, after determining that the service data is abnormal, an alarm is issued to indicate that the abnormal service data occurs to the user, and a specific alarm method may be voice broadcast, and the like, which is not specifically limited herein.

Based on the same technical concept, fig. 7 exemplarily shows a schematic structural diagram of a service data monitoring device according to an embodiment of the present invention, and the device can execute a flow of a service data monitoring method.

As shown in fig. 7, the apparatus specifically includes:

an obtaining module 710, configured to obtain first service data in a preset time period and second service data in a historical time period associated with the preset time period; the first service data and the second service data aim at the same service index;

a processing module 720, configured to generate an image to be identified, which represents the first service data and the second service data;

Optionally, the processing module 720 is specifically configured to:

and preprocessing the graph to determine the image to be identified.

Optionally, the processing module 720 is specifically configured to:

generating a picture by the graph;

Optionally, the processing module 720 is further configured to:

the processing module 720 is specifically configured to:

Optionally, the processing module 720 is further configured to:

Based on the same technical concept, an embodiment of the present invention further provides a computer device, including:

a memory for storing program instructions;

Based on the same technical concept, an embodiment of the present invention further provides a computer-readable storage medium, where computer-executable instructions are stored, and the computer-executable instructions are used to enable a computer to execute the method for monitoring the service data.

As will be appreciated by one skilled in the art, embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.

The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems), and computer program products according to the application. It will be understood that each flow and/or block of the flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.

These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.

It will be apparent to those skilled in the art that various changes and modifications may be made in the present application without departing from the spirit and scope of the application. Thus, if such modifications and variations of the present application fall within the scope of the claims of the present application and their equivalents, the present application is intended to include such modifications and variations as well.

Claims

1. A method for monitoring service data is characterized by comprising the following steps:

2. The method of claim 1, wherein the traffic indicator is at least one of: the service transaction amount, the average service consumption time and the service success rate;

3. The method of claim 1, wherein generating the image to be identified characterizing the first traffic data and the second traffic data comprises:

and preprocessing the graph to determine the image to be identified.

4. The method of claim 1, wherein pre-processing the graph to determine the image to be identified comprises:

generating a picture by the graph;

5. The method of claim 1, wherein the inputting the image to be recognized into a convolutional neural network model, and after determining a monitoring result of the traffic data in the preset time period under the traffic index, further comprises:

6. The method of claim 1, wherein the convolutional neural network model is used to perform N classification;

7. The method of any of claims 1 to 6, further comprising:

8. A device for monitoring traffic data, comprising:

9. A computer device, comprising:

a memory for storing program instructions;

a processor for calling program instructions stored in said memory to perform the method of any of claims 1 to 7 in accordance with the obtained program.

10. A computer-readable storage medium having stored thereon computer-executable instructions for causing a computer to perform the method of any one of claims 1 to 7.