CN113360816A - Click rate prediction method and device - Google Patents
Click rate prediction method and device Download PDFInfo
- Publication number
- CN113360816A CN113360816A CN202010147662.7A CN202010147662A CN113360816A CN 113360816 A CN113360816 A CN 113360816A CN 202010147662 A CN202010147662 A CN 202010147662A CN 113360816 A CN113360816 A CN 113360816A
- Authority
- CN
- China
- Prior art keywords
- user
- historical
- behavior
- sequence
- item
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 41
- 239000013598 vector Substances 0.000 claims abstract description 44
- 230000006399 behavior Effects 0.000 claims description 255
- 238000004590 computer program Methods 0.000 claims description 10
- 238000013528 artificial neural network Methods 0.000 claims description 9
- 230000007246 mechanism Effects 0.000 claims description 8
- 230000007613 environmental effect Effects 0.000 abstract description 6
- 238000010586 diagram Methods 0.000 description 12
- 230000006870 function Effects 0.000 description 9
- 238000004891 communication Methods 0.000 description 7
- 238000012545 processing Methods 0.000 description 6
- 230000003287 optical effect Effects 0.000 description 4
- 230000008569 process Effects 0.000 description 4
- 230000009471 action Effects 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 239000000835 fiber Substances 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 230000000644 propagated effect Effects 0.000 description 2
- 239000004065 semiconductor Substances 0.000 description 2
- 230000003542 behavioural effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 239000004973 liquid crystal related substance Substances 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/02—Marketing; Price estimation or determination; Fundraising
- G06Q30/0201—Market modelling; Market analysis; Collecting market data
- G06Q30/0202—Market predictions or forecasting for commercial activities
Landscapes
- Engineering & Computer Science (AREA)
- Business, Economics & Management (AREA)
- Accounting & Taxation (AREA)
- Development Economics (AREA)
- Finance (AREA)
- Strategic Management (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Entrepreneurship & Innovation (AREA)
- Databases & Information Systems (AREA)
- Physics & Mathematics (AREA)
- Game Theory and Decision Science (AREA)
- General Business, Economics & Management (AREA)
- Marketing (AREA)
- Economics (AREA)
- General Engineering & Computer Science (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a click rate prediction method and device, and relates to the technical field of computers. One embodiment of the method comprises: determining a microscopic behavior sequence of the user based on historical behavior data corresponding to the user; splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector; predicting the click rate of the user on the candidate item based on the user behavior feature vector. According to the implementation mode, the click rate is predicted by adopting the microscopic behavior sequence capable of reflecting the interest and behavior habits of the user and the portrait characteristics and environmental characteristics of the user, so that the accuracy of the prediction result can be improved.
Description
Technical Field
The invention relates to the technical field of computers, in particular to a click rate prediction method and device.
Background
Click-through Rate (CTR) is an important measure in industrial applications. In the conventional technology, when the click rate is predicted, the prediction is generally performed based on the macro behavior data of the user, namely, the item information related to the past historical data.
However, the macro behavior data cannot accurately reflect the interests and behavior habits of the user, and the accuracy of predicting the click rate based on the macro behavior data is poor.
Disclosure of Invention
In view of this, embodiments of the present invention provide a method and an apparatus for predicting a click rate, which are capable of improving accuracy of a prediction result by using a micro behavior sequence capable of reflecting interests and behavior habits of a user and an image characteristic and an environmental characteristic of the user to predict the click rate.
To achieve the above object, according to an aspect of an embodiment of the present invention, there is provided a click rate prediction method, including:
determining a microscopic behavior sequence of the user based on historical behavior data corresponding to the user;
splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector;
predicting the click rate of the user on the candidate item based on the user behavior feature vector.
Optionally, determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user includes:
dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
Optionally, the microscopic behavior sequence comprises: a first sequence of behaviors and a second sequence of behaviors;
determining a microscopic behavior sequence of each historical item within the any session, comprising: for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session; generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session; and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
Optionally, the association comprises at least one of: the historical item belongs to the category, the brand of the historical item, the shop where the historical item is located and the keyword of the historical item.
Optionally, after obtaining the micro-behavior sequence of any historical item in any session, the method further includes:
determining an attention weight of the any historical item based on an attention mechanism and the word vector of the candidate item; adjusting a microscopic behavior sequence of the any historical item within the any session based on the attention weight of the any historical item.
Optionally, predicting the click rate of the user on the candidate item based on the user behavior feature vector includes: and predicting the click rate of the user on the candidate item by adopting a full-connection neural network based on the user behavior feature vector.
Optionally, before determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user, the method further includes:
judging whether the user has historical behavior data or not;
if yes, judging whether the historical behavior data of the user meets a preset requirement or not; if so, taking the historical data of the user as the historical data corresponding to the user; if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user;
and if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user.
According to a second aspect of the embodiments of the present invention, there is provided an apparatus for predicting a click rate, including:
the determining module is used for determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user;
the splicing module is used for splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector;
and the prediction module predicts the click rate of the user on the candidate item based on the user behavior feature vector.
Optionally, the determining module determines the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user, including:
dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
Optionally, the microscopic behavior sequence comprises: a first sequence of behaviors and a second sequence of behaviors;
the determining module determines a microscopic behavior sequence of each historical item within the any session, including: for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session; generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session; and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
Optionally, the association comprises at least one of: the historical item belongs to the category, the brand of the historical item, the shop where the historical item is located and the keyword of the historical item.
Optionally, the determining module is further configured to: after obtaining the microscopic behavior sequence of any historical item in any session, determining the attention weight of any historical item based on an attention mechanism and the word vector of the candidate item; adjusting a microscopic behavior sequence of the any historical item within the any session based on the attention weight of the any historical item.
Optionally, the determining module is further configured to: before determining a microscopic behavior sequence of a user based on historical behavior data corresponding to the user, judging whether the user has the historical behavior data;
if yes, judging whether the historical behavior data of the user meets a preset requirement or not; if so, taking the historical data of the user as the historical data corresponding to the user; if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user;
and if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user.
Optionally, the predicting module predicts the click rate of the user on the candidate item based on the user behavior feature vector, including: and predicting the click rate of the user on the candidate item by adopting a full-connection neural network based on the user behavior feature vector.
According to a third aspect of the embodiments of the present invention, there is provided an electronic device for click rate prediction, including:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method provided by the first aspect of the embodiments of the present invention.
According to a fourth aspect of embodiments of the present invention, there is provided a computer readable medium, on which a computer program is stored, which when executed by a processor, implements the method provided by the first aspect of embodiments of the present invention.
One embodiment of the above invention has the following advantages or benefits: the click rate is predicted by adopting the microscopic behavior sequence capable of reflecting the interest and behavior habits of the user and the portrait characteristics and environmental characteristics of the user, so that the accuracy of the prediction result can be improved. The behavior data in the user group where the user is located is used as the historical behavior data corresponding to the user, and the problem of cold start of a new user can be solved. The click rate prediction is carried out by using historical behavior data in one session, so that the accuracy of the prediction result can be further improved.
Further effects of the above-mentioned non-conventional alternatives will be described below in connection with the embodiments.
Drawings
The drawings are included to provide a better understanding of the invention and are not to be construed as unduly limiting the invention. Wherein:
FIG. 1 is a schematic diagram of a main flow of a method of click-through rate prediction according to an embodiment of the present invention;
FIG. 2 is a schematic diagram of a method of click rate prediction in an alternative embodiment of the invention;
FIG. 3 is a schematic diagram of the main blocks of an apparatus for click-through rate prediction according to an embodiment of the present invention;
FIG. 4 is an exemplary system architecture diagram in which embodiments of the present invention may be employed;
fig. 5 is a schematic block diagram of a computer system suitable for use in implementing a terminal device or server of an embodiment of the invention.
Detailed Description
Exemplary embodiments of the present invention are described below with reference to the accompanying drawings, in which various details of embodiments of the invention are included to assist understanding, and which are to be considered as merely exemplary. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the invention. Also, descriptions of well-known functions and constructions are omitted in the following description for clarity and conciseness.
According to an aspect of an embodiment of the present invention, a method for click rate prediction is provided.
Fig. 1 is a schematic diagram of a main flow of a click-through rate prediction method according to an embodiment of the present invention, and as shown in fig. 1, the click-through rate prediction method includes: step S101, step S102, and step S103.
And S101, determining a microscopic behavior sequence of the user based on historical behavior data corresponding to the user.
The macro behavior data refers to the article information in the past historical data of the user. A macro behavior will typically contain information about a variety of micro behaviors. For example, in the e-commerce field, how the user enters the merchandise page, whether to browse the detailed description of the merchandise, whether to join a shopping cart or place an order, whether to recently purchase or purchase the merchandise of the same third-level category, whether to recently purchase or purchase the merchandise of the same brand, whether to recently purchase or purchase the merchandise of the same product word, whether to recently purchase or add the merchandise of the same shop, and the like. Micro behavior data is relative to macro behavior data and refers to data that is finer in granularity than macro behavior data. The microscopic behavior sequence refers to a sequence composed of a plurality of items of microscopic behavior data.
In the existing deep learning method, the click rate is generally predicted by macroscopic behavior data of a user. However, the macro behavior data is easy to include other operations such as wrong clicks, and the interests and behavior habits of the user cannot be accurately reflected. According to the invention, the click rate is predicted by adopting the microscopic behavior data capable of reflecting the interests and behavior habits of the user, and the accuracy of the prediction result can be improved.
Optionally, determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user includes: dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
The dividing manner of the session may be selectively set according to actual situations, for example, taking 30 minutes or one hour as a dividing unit, and dividing the data in each dividing unit in the historical behavior data into one session. User behavior within one session is generally more similar, with no strong correlation between behavior within different sessions. By dividing the session, the accuracy of the prediction result can be further improved.
In the practical application process, the number of the articles in one session can be limited. For example, an upper limit on the number of items may be limited for active users, such as containing up to 200 items; the number of the items can be limited for the inactive users, for example, the conversation division is performed according to the action sequence which takes the latest time. Therefore, the uniformity and the effectiveness of data in each session can be ensured, and the accuracy of a prediction result is improved.
In this example, the microscopic behavior sequences of the plurality of historical articles are added, so that the lengths of the microscopic behavior sequences corresponding to different sessions are equal, and the subsequent analysis and processing are facilitated.
Optionally, the microscopic behavior sequence comprises: a first sequence of behaviors and a second sequence of behaviors. Determining a microscopic behavior sequence of each historical item within the any session, comprising: for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session; generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session; and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
The associated item refers to a microscopic information item related to the historical item. Optionally, the association comprises at least one of: the historical item belongs to the category, the brand of the historical item, the shop where the historical item is located and the keyword of the historical item.
The behavioral interest of the user can be more accurately represented through microscopic behaviors, and in a certain time range, the longer the user stays in a certain article, the greater the interest in the article, and the same reason is that information browsing or operation information is obtained. Meanwhile, if the user has more recent behaviors on the brand or the category, the user can be considered to be more interested in the brand or the category, and the shop and the keyword information are the same.
The source channel refers to a channel through which a user acquires an article, and examples of the source channel include a search page (search), a shopping cart page, a promotion page (promotion), a homepage, and the like. The browsing information content may include header information (item title, price, and item host), comments, item details (detail), and the like. The operation information refers to an operation action of the user on the article, such as adding a shopping cart (addCart), paying attention, purchasing or placing an order (order), and the like. The stay time is used for reflecting the stay time of the user on the article, and the stay time can be divided into a plurality of time intervals in advance in the practical application process, and the time interval to which the stay time of the user on the article belongs is taken as the stay time. In practical application, a positive number can uniquely represent the content of one item of microscopic information, and 0 represents that the content of the corresponding microscopic information item is empty. Take the source channel as an example. Assume that there are four source channels: search page, shopping cart page, promotion page, and home page, the search page, shopping cart page, promotion page, and home page may be represented by numbers 1, 2, 3, and 4, respectively.
Illustratively, the macro-behavior sequence of a certain user in a session is (item 1, item 2, item 3). The first behavior sequences of product 1, product 2, and product 3 are: (1,3,1,5),(3,0,3,7),(3,3,0,4). The second behavior sequence is the preference degree of the user for the same category, brand, shop and product word of the goods in the browsing sequence, such as the number of the same category goods browsed or bought by the user in the last 7 days, and the brand, shop and product word information is obtained similarly, and is specifically represented as ((3, 7, 10, 5), (4, 2, 1, 7), (3, 5, 2, 4)). And splicing the first row sequence and the second row sequence of each commodity, wherein the micro row sequence of the commodity 1 in the current session is (1, 3, 1, 5, 3, 7, 10, 5), the micro row sequence of the commodity 2 is (3, 0, 3, 7, 4, 2, 1, 7), and the micro row sequence of the commodity 3 is (3, 3, 0, 4, 3, 5, 2, 4). Adding the microscopic behavior sequences of each commodity in the current conversation to obtain the microscopic behavior sequence of the user in the current conversation as follows: (7,6,4, 16, 10, 14, 13, 16)
Optionally, after obtaining the micro-behavior sequence of any historical item in any session, the method further includes: determining an Attention weight of the any historical item based on an Attention Mechanism (Attention Mechanism, a Mechanism for improving model effect) and a word vector of the candidate item; adjusting a microscopic behavior sequence of the any historical item within the any session based on the attention weight of the any historical item. The word vector of the candidate item is the same dimension as the microscopic behavior sequence of the historical item. For example, the microscopic behavior sequence of the historical item includes a first behavior sequence and a second behavior sequence, and the word vector of the candidate item also includes the first behavior sequence and the second behavior sequence.
For example, if the user has recently browsed lipstick-type products and has an additional purchase behavior, it is easier to arouse the interest of the user to generate a click behavior when the candidate product is lipstick, so that a higher weight needs to be set for the lipstick-type products in the user historical behavior data. The specific method comprises the following steps:
wherein (e)1,e2,…,en) Sequence of microscopic behaviors for individual historical items within a conversation, vcIs a word vector of the candidate commodity. attention () is a kind of feedforward neural network used to calculate attention weight of candidate item and user's microscopic behavior sequence. Vu(C) And C, the user micro behavior sequence is the adjusted user micro behavior sequence when the candidate object is C. The user can have different vector representations for different candidate articles (articles to be predicted) by adjusting the microscopic behavior sequence of the user based on the attention mechanism.
And S102, splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector.
The user portrait characteristics refer to various items of information of the user, and may include attribute characteristics of the user, such as the gender of the user, province identification, whether a car exists, whether a pet exists, and the like. The environmental characteristics refer to interaction information between the user and the system related to the candidate item, and may specifically include information such as browsing, purchase adding, order placing and the like of the item to which the candidate item belongs in the last three months of the user, or preference degree information and the like of the item.
And S103, predicting the click rate of the user on the candidate item based on the user behavior feature vector.
The closer the user behavior feature vector is to the word vector of the candidate item, the greater the probability of the user clicking on the candidate item. Optionally, as shown in fig. 2, predicting the click rate of the candidate item by the user based on the user behavior feature vector includes: and predicting the click rate of the user on the candidate item by adopting a full-connection neural network based on the user behavior feature vector. In fig. 2, each session corresponds to one user behavior feature vector, and a circle in each user behavior feature vector corresponds to one item. The fully connected layer can learn the interaction among the behavior feature vectors of the users. In practical application, the negative log-likelihood function of the following formula can be used as the loss function of the network:
wherein L represents a negative likelihood function; x represents a user behavior feature vector and is input into the neural network; y is the corresponding label, and is determined by whether the article is clicked, if the article is clicked, y is 1, otherwise, y is 0. p (x) represents the output of the neural network after passing through the full connection layer (softmax), namely the probability that the candidate commodity is clicked is predicted by the neural network; and N is the number of the user behavior feature vectors.
Optionally, before determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user, the method further includes:
judging whether the user has historical behavior data or not;
if yes, judging whether the historical behavior data of the user meets a preset requirement or not; if so, taking the historical data of the user as the historical data corresponding to the user; if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user;
and if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user.
The preset requirement may be selectively set according to an actual situation, for example, the data amount of the historical behavior data is not less than a preset data amount threshold, the time span of the historical behavior data is not less than a preset time span threshold, and the like. In this example, when the user has no historical behavior data or the historical behavior data does not meet the preset requirement, the behavior data in the user group where the user is located is used as the historical behavior data corresponding to the user, so that the problem of cold start of a new user can be solved.
The click rate prediction method can be applied to commodity recommendation in the E-commerce field, and candidate commodities with high predicted values are recommended to users by predicting the click rates of the candidate commodities.
According to a second aspect of the embodiments of the present invention, there is provided an apparatus for implementing the above method.
FIG. 3 is a schematic diagram of the main blocks of an apparatus for click rate prediction according to an embodiment of the present invention, and as shown in FIG. 3, the apparatus 300 for click rate prediction includes:
the determining module 301 determines a microscopic behavior sequence of the user based on historical behavior data corresponding to the user;
the splicing module 302 is used for splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector;
and the prediction module 303 predicts the click rate of the candidate item by the user based on the user behavior feature vector.
Optionally, the determining module determines the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user, including:
dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
Optionally, the microscopic behavior sequence comprises: a first sequence of behaviors and a second sequence of behaviors;
the determining module determines a microscopic behavior sequence of each historical item within the any session, including: for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session; generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session; and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
Optionally, the association comprises at least one of: the historical item belongs to the category, the brand of the historical item, the shop where the historical item is located and the keyword of the historical item.
Optionally, the determining module is further configured to: after obtaining the microscopic behavior sequence of any historical item in any session, determining the attention weight of any historical item based on an attention mechanism and the word vector of the candidate item; adjusting a microscopic behavior sequence of the any historical item within the any session based on the attention weight of the any historical item.
Optionally, the determining module is further configured to: before determining a microscopic behavior sequence of a user based on historical behavior data corresponding to the user, judging whether the user has the historical behavior data;
if yes, judging whether the historical behavior data of the user meets a preset requirement or not; if so, taking the historical data of the user as the historical data corresponding to the user; if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user;
and if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user.
Optionally, the predicting module predicts the click rate of the user on the candidate item based on the user behavior feature vector, including: and predicting the click rate of the user on the candidate item by adopting a full-connection neural network based on the user behavior feature vector.
According to a third aspect of the embodiments of the present invention, there is provided an electronic device for click rate prediction, including:
one or more processors;
a storage device for storing one or more programs,
when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement the method provided by the first aspect of the embodiments of the present invention.
According to a fourth aspect of embodiments of the present invention, there is provided a computer readable medium, on which a computer program is stored, which when executed by a processor, implements the method provided by the first aspect of embodiments of the present invention.
FIG. 4 illustrates an exemplary system architecture 400 of a method of click-through rate prediction or a device of click-through rate prediction to which embodiments of the present invention may be applied.
As shown in fig. 4, the system architecture 400 may include terminal devices 401, 402, 403, a network 404, and a server 405. The network 404 serves as a medium for providing communication links between the terminal devices 401, 402, 403 and the server 405. Network 404 may include various types of connections, such as wire, wireless communication links, or fiber optic cables, to name a few.
A user may use terminal devices 401, 402, 403 to interact with a server 405 over a network 404 to receive or send messages or the like. The terminal devices 401, 402, 403 may have installed thereon various communication client applications, such as shopping-like applications, web browser applications, search-like applications, instant messaging tools, mailbox clients, social platform software, etc. (by way of example only).
The terminal devices 401, 402, 403 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop portable computers, desktop computers, and the like.
The server 405 may be a server providing various services, such as a background management server (for example only) providing support for shopping websites browsed by users using the terminal devices 401, 402, 403. The backend management server may analyze and perform other processing on the received data such as the product information query request, and feed back a processing result (for example, target push information, product information — just an example) to the terminal device.
It should be noted that the method for predicting the click rate provided by the embodiment of the present invention is generally executed by the server 405, and accordingly, the device for predicting the click rate is generally disposed in the server 405.
It should be understood that the number of terminal devices, networks, and servers in fig. 4 is merely illustrative. There may be any number of terminal devices, networks, and servers, as desired for implementation.
Referring now to FIG. 5, shown is a block diagram of a computer system 500 suitable for use with a terminal device implementing an embodiment of the present invention. The terminal device shown in fig. 5 is only an example, and should not bring any limitation to the functions and the scope of use of the embodiments of the present invention.
As shown in fig. 5, the computer system 500 includes a Central Processing Unit (CPU)501 that can perform various appropriate actions and processes according to a program stored in a Read Only Memory (ROM)502 or a program loaded from a storage section 508 into a Random Access Memory (RAM) 503. In the RAM 503, various programs and data necessary for the operation of the system 500 are also stored. The CPU 501, ROM 502, and RAM 503 are connected to each other via a bus 504. An input/output (I/O) interface 505 is also connected to bus 504.
The following components are connected to the I/O interface 505: an input portion 506 including a keyboard, a mouse, and the like; an output portion 507 including a display such as a Cathode Ray Tube (CRT), a Liquid Crystal Display (LCD), and the like, and a speaker; a storage portion 508 including a hard disk and the like; and a communication section 509 including a network interface card such as a LAN card, a modem, or the like. The communication section 509 performs communication processing via a network such as the internet. The driver 510 is also connected to the I/O interface 505 as necessary. A removable medium 511 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, or the like is mounted on the drive 510 as necessary, so that a computer program read out therefrom is mounted into the storage section 508 as necessary.
In particular, according to the embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, embodiments of the present disclosure include a computer program product comprising a computer program embodied on a computer readable medium, the computer program comprising program code for performing the method illustrated in the flow chart. In such an embodiment, the computer program may be downloaded and installed from a network through the communication section 509, and/or installed from the removable medium 511. The computer program performs the above-described functions defined in the system of the present invention when executed by the Central Processing Unit (CPU) 501.
It should be noted that the computer readable medium shown in the present invention can be a computer readable signal medium or a computer readable storage medium or any combination of the two. A computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the foregoing. More specific examples of the computer readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present invention, a computer readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. In the present invention, however, a computer readable signal medium may include a propagated data signal with computer readable program code embodied therein, for example, in baseband or as part of a carrier wave. Such a propagated data signal may take many forms, including, but not limited to, electro-magnetic, optical, or any suitable combination thereof. A computer readable signal medium may also be any computer readable medium that is not a computer readable storage medium and that can communicate, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, fiber optic cable, RF, etc., or any suitable combination of the foregoing.
The flowchart and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present invention. In this regard, each block in the flowchart or block diagrams may represent a module, segment, or portion of code, which comprises one or more executable instructions for implementing the specified logical function(s). It should also be noted that, in some alternative implementations, the functions noted in the block may occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams or flowchart illustration, and combinations of blocks in the block diagrams or flowchart illustration, can be implemented by special purpose hardware-based systems which perform the specified functions or acts, or combinations of special purpose hardware and computer instructions.
The modules described in the embodiments of the present invention may be implemented by software or hardware. The described modules may also be provided in a processor, which may be described as: a processor comprising: the determining module is used for determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user; the splicing module is used for splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector; and the prediction module predicts the click rate of the user on the candidate item based on the user behavior feature vector. Where the names of these modules do not in some cases constitute a limitation of the module itself, for example, a determination module may also be described as a "module that stitches the sequence of microscopic behaviors with the portrait and environmental features of the user".
As another aspect, the present invention also provides a computer-readable medium that may be contained in the apparatus described in the above embodiments; or may be separate and not incorporated into the device. The computer readable medium carries one or more programs which, when executed by a device, cause the device to comprise: determining a microscopic behavior sequence of the user based on historical behavior data corresponding to the user; splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector; predicting the click rate of the user on the candidate item based on the user behavior feature vector.
According to the technical scheme of the embodiment of the invention, the accuracy of the prediction result can be improved by adopting the micro behavior sequence capable of reflecting the interest and behavior habits of the user and the prediction click rate of the portrait characteristics and the environmental characteristics of the user. The behavior data in the user group where the user is located is used as the historical behavior data corresponding to the user, and the problem of cold start of a new user can be solved. The click rate prediction is carried out by using historical behavior data in one session, so that the accuracy of the prediction result can be further improved.
The above-described embodiments should not be construed as limiting the scope of the invention. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions can occur, depending on design requirements and other factors. Any modification, equivalent replacement, and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (12)
1. A method of click rate prediction, comprising:
determining a microscopic behavior sequence of the user based on historical behavior data corresponding to the user;
splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector;
predicting the click rate of the user on the candidate item based on the user behavior feature vector.
2. The method of claim 1, wherein determining a sequence of microscopic behaviors of the user based on historical behavior data corresponding to the user comprises:
dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
3. The method of claim 2, wherein the sequence of microscopic behaviors comprises: a first sequence of behaviors and a second sequence of behaviors;
determining a microscopic behavior sequence of each historical item within the any session, comprising:
for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session;
generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session;
and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
4. The method of claim 3, wherein the association comprises at least one of: the historical item belongs to the category, the brand of the historical item, the shop where the historical item is located and the keyword of the historical item.
5. The method of claim 3, wherein after obtaining the sequence of microscopic behaviors of the any historical item within the any session, further comprising:
determining an attention weight of the any historical item based on an attention mechanism and the word vector of the candidate item; adjusting a microscopic behavior sequence of the any historical item within the any session based on the attention weight of the any historical item.
6. The method of claim 1, wherein predicting the user's click-through rate for candidate items based on the user behavior feature vector comprises: and predicting the click rate of the user on the candidate item by adopting a full-connection neural network based on the user behavior feature vector.
7. The method of claim 1, wherein prior to determining the sequence of microscopic behaviors of the user based on the historical behavior data corresponding to the user, further comprising:
judging whether the user has historical behavior data or not;
if yes, judging whether the historical behavior data of the user meets a preset requirement or not; if so, taking the historical data of the user as the historical data corresponding to the user; if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user;
and if not, taking the behavior data in the user group where the user is as the historical behavior data corresponding to the user.
8. An apparatus for click rate prediction, comprising:
the determining module is used for determining the microscopic behavior sequence of the user based on the historical behavior data corresponding to the user;
the splicing module is used for splicing the microscopic behavior sequence with the portrait characteristics and the environment characteristics of the user to obtain a user behavior characteristic vector;
and the prediction module predicts the click rate of the user on the candidate item based on the user behavior feature vector.
9. The apparatus of claim 8, wherein the prediction module determines a microscopic behavior sequence of the user based on historical behavior data corresponding to the user, comprising:
dividing historical behavior data corresponding to the user into a plurality of sessions; determining a microscopic behavior sequence of each historical item within any one session; and adding the microscopic behavior sequences of each historical item to obtain the microscopic behavior sequence of the user in any session.
10. The apparatus of claim 9, wherein the sequence of microscopic behaviors comprises: a first sequence of behaviors and a second sequence of behaviors;
the prediction module determines a microscopic behavior sequence of each historical item within the any session, including:
for any historical item, generating a first behavior sequence of the historical item in any session according to a source channel, browsing information, operation information and stay time of the historical item in any session;
generating a second behavior sequence of any historical item in any session according to the preference degree of the user on the associated item of any historical item in any session;
and splicing the first behavior sequence and the second behavior sequence to obtain the microscopic behavior sequence of any historical article in any session.
11. An electronic device for click through rate prediction, comprising:
one or more processors;
a storage device for storing one or more programs,
when executed by the one or more processors, cause the one or more processors to implement the method of any one of claims 1-7.
12. A computer-readable medium, on which a computer program is stored, which, when being executed by a processor, carries out the method according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010147662.7A CN113360816A (en) | 2020-03-05 | 2020-03-05 | Click rate prediction method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010147662.7A CN113360816A (en) | 2020-03-05 | 2020-03-05 | Click rate prediction method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113360816A true CN113360816A (en) | 2021-09-07 |
Family
ID=77523822
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010147662.7A Pending CN113360816A (en) | 2020-03-05 | 2020-03-05 | Click rate prediction method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113360816A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113744030A (en) * | 2021-09-08 | 2021-12-03 | 未鲲(上海)科技服务有限公司 | Recommendation method, device, server and medium based on AI user portrait |
| CN113962753A (en) * | 2021-12-22 | 2022-01-21 | 浙江口碑网络技术有限公司 | Candidate commodity ordering method, candidate commodity display method, candidate commodity ordering device, candidate commodity display device and electronic equipment |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150019373A1 (en) * | 2013-07-09 | 2015-01-15 | International Business Machines Corporation | Providing a consumer advocate recommendation utilizing historic purchasing data |
| CN109960759A (en) * | 2019-03-22 | 2019-07-02 | 中山大学 | Prediction method of click-through rate of recommendation system based on deep neural network |
| CN110807156A (en) * | 2019-10-23 | 2020-02-18 | 山东师范大学 | An interest recommendation method and system based on user sequence click behavior |
-
2020
- 2020-03-05 CN CN202010147662.7A patent/CN113360816A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20150019373A1 (en) * | 2013-07-09 | 2015-01-15 | International Business Machines Corporation | Providing a consumer advocate recommendation utilizing historic purchasing data |
| CN109960759A (en) * | 2019-03-22 | 2019-07-02 | 中山大学 | Prediction method of click-through rate of recommendation system based on deep neural network |
| CN110807156A (en) * | 2019-10-23 | 2020-02-18 | 山东师范大学 | An interest recommendation method and system based on user sequence click behavior |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113744030A (en) * | 2021-09-08 | 2021-12-03 | 未鲲(上海)科技服务有限公司 | Recommendation method, device, server and medium based on AI user portrait |
| CN113962753A (en) * | 2021-12-22 | 2022-01-21 | 浙江口碑网络技术有限公司 | Candidate commodity ordering method, candidate commodity display method, candidate commodity ordering device, candidate commodity display device and electronic equipment |
| CN113962753B (en) * | 2021-12-22 | 2022-05-13 | 浙江口碑网络技术有限公司 | Candidate commodity ordering method, candidate commodity display method, candidate commodity ordering device, candidate commodity display device and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US20200250732A1 (en) | Method and apparatus for use in determining tags of interest to user | |
| JP5899275B2 (en) | System and method for scoring quality of advertisement and content in online system | |
| JP5646787B1 (en) | Integrated market for advertising and content in online systems | |
| US12141212B2 (en) | Intelligent interface accelerating | |
| CN111095330B (en) | Machine learning method and system for predicting online user interactions | |
| CN110020162B (en) | User identification method and device | |
| CN113450172B (en) | Commodity recommendation method and device | |
| CN113095893A (en) | Method and device for determining sales of articles | |
| CN109961299A (en) | The method and apparatus of data analysis | |
| CN111311294A (en) | Data processing method, apparatus, medium and electronic equipment | |
| CN108197298A (en) | A kind of smart shopper exchange method and system based on natural language processing | |
| WO2023142520A1 (en) | Information recommendation method and apparatus | |
| CN113495991A (en) | Recommendation method and device | |
| CN113793161B (en) | Advertisement delivery method, advertisement delivery device, readable storage medium and electronic equipment | |
| CN113139113A (en) | Search request processing method and device | |
| CN112749323B (en) | Method and device for constructing user portrait | |
| CN113360816A (en) | Click rate prediction method and device | |
| CN114936873B (en) | E-commerce traffic conversion intelligent guidance analysis method, device, storage medium and electronic device | |
| CN113822734B (en) | Method and device for generating information | |
| CN110232581B (en) | Method and device for providing coupons for users | |
| CN113269600B (en) | Information sending method and device | |
| CN113159877B (en) | Data processing method, device, system and computer readable storage medium | |
| CN113763112A (en) | Information pushing method and device | |
| US20160275535A1 (en) | Centralized system for progressive price management | |
| CN117573973A (en) | Resource recommendation method, device, electronic equipment and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |