Click rate estimation method and device based on click model, electronic equipment and medium

Abstract

The embodiment of the application provides a click rate estimation method and device based on a click model, electronic equipment and medium, and belongs to the technical field of data processing. The method comprises the following steps: obtaining historical behavior data of a user clicking a browser within a preset historical time length; inputting the historical behavior data into a click model for differential evolution to obtain a test vector; determining a target vector according to the test vector and the historical behavior data; training the click model according to the target vector to obtain a click estimation model; predicting the click rate of the click behavior of the user according to the click estimation model, and outputting demand information corresponding to the click behavior; and recommending a display page corresponding to the demand information to the user. The embodiment of the application can avoid falling into a local optimum in the prediction process, can recommend a corresponding display page to the user according to the demand of the user, and realizes accurate estimation of the click rate.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of data processing, and particularly relates to a click rate estimation method and device based on a click model, an electronic device and a medium. BACKGROUND

[0002] With the development of information technology, people can obtain information they want to know through online browsing, such as insurance information, medical information, and the like. The rise of the Internet enables people to see different media content when browsing the same page, thereby realizing personalized display of media content. For example, in the process of webpage test optimization, testers test the click rate of visitors to understand media content that different users are interested in, such as user-intended insurance types and user-intended investment types, so as to more accurately display corresponding media content to each user, thereby improving the click rate of media content and improving the access volume of the page.

[0003] However, the click model may have a large number of model gradient descent iteration times, resulting in slow convergence speed and low prediction accuracy, thereby causing deviation in click rate prediction and reducing prediction accuracy, and thus reducing the user experience of browsing the page. SUMMARY

[0004] The main purpose of the embodiments of the present application is to provide a click rate estimation method and device based on a click model, an electronic device and a medium, which can recommend corresponding display pages to users according to user needs, and accurately estimate the click rate.

[0005] To achieve the above purpose, a first aspect of the embodiments of the present application provides a click rate estimation method based on a click model, which comprises:

[0006] acquiring historical behavior data of a user clicking a browser in a preset historical time length;

[0007] inputting the historical behavior data into the click model for differential evolution to obtain a test vector;

[0008] determining a target vector according to the test vector and the historical behavior data;

[0009] training the click model according to the target vector to output demand information corresponding to the click behavior;

[0010] recommending a display page corresponding to the demand information to the user.

[0011] In some embodiments, the inputting the historical behavior data into the click model for differential evolution calculation to obtain a test vector comprises:

[0012] inputting the historical behavior data into the click model to perform an initialization calculation, to obtain an initialization vector and an optimal individual vector of the historical behavior data;

[0013] performing a mutation operation on the historical behavior data according to the optimal individual vector and a predetermined crossover vector, to obtain a mutation vector;

[0014] performing a crossover operation on the mutation vector and the initialization vector, to obtain the test vector.

[0015] In some embodiments, the performing of the mutation operation on the historical behavior data according to the optimal individual vector and the predetermined crossover vector, to obtain a mutation vector, comprises:

[0016] obtaining a current individual vector of the historical behavior data;

[0017] determining an optimal difference vector according to the current individual vector and the optimal individual vector;

[0018] determining an iteration difference vector according to the crossover vector and the current individual vector;

[0019] performing a mutation operation based on a preset scaling factor, the optimal difference vector, the iteration difference vector and the current individual vector, to obtain the mutation vector.

[0020] In some embodiments, the performing of the crossover operation on the mutation vector and the initialization vector, to obtain the test vector, comprises:

[0021] performing a binomial crossover operation on the mutation vector and the initialization vector based on a preset crossover probability, to obtain the test vector.

[0022] In some embodiments, the determining of the target vector according to the test vector and the historical behavior data comprises:

[0023] determining a plurality of historical behavior vectors according to the historical behavior data;

[0024] performing fitness calculation on the historical behavior vectors, to obtain historical fitness values of the historical behavior vectors;

[0025] performing fitness calculation on the test vector, to obtain a test fitness value of the test vector;

[0026] comparing the historical fitness values and the test fitness value, to obtain a comparison result;

[0027] determining the target vector from the historical behavior vectors and the test vector according to the comparison result.

[0028] In some embodiments, the click model comprises a back propagation neural network, and the training the click model according to the target vector comprises:

[0029] performing gradient descent calculation on the back propagation neural network to obtain network weights and network thresholds of the back propagation neural network;

[0030] training the network weights and the network thresholds according to the target vector to obtain a fitness function;

[0031] training the click model according to the fitness function to obtain the click prediction model.

[0032] In some embodiments, after the training the click model according to the target vector to obtain the click prediction model, the method further comprises:

[0033] updating the initial vector every preset optimization times.

[0034] To achieve the above object, a second aspect of the embodiments of the present application provides a click rate prediction device based on a click model, the device comprising:

[0035] a data acquisition module configured to acquire historical behavior data of a user clicking a browser in a preset historical time period;

[0036] a differential evolution module configured to input the historical behavior data into the click model to perform differential evolution calculation to obtain a trial vector;

[0037] a vector determination module configured to determine a target vector according to the trial vector and the historical behavior data;

[0038] a model training module configured to train the click model according to the target vector to obtain a click prediction model;

[0039] a data prediction module configured to predict a click rate of a click behavior of a user according to the click prediction model, and output demand information corresponding to the click behavior;

[0040] a page display module configured to recommend a display page corresponding to the demand information to the user.

[0041] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the click rate prediction method based on a click model when executing the computer program.

[0042] To achieve the above object, a fourth aspect of the embodiment of the present application provides a computer readable storage medium, the storage medium stores a computer program, and the computer program is executed by a processor to implement the click rate estimation method based on a click model.

[0043] The click rate estimation method, device, electronic equipment and storage medium based on a click model provided by the present application first acquire historical behavior data of a user clicking a browser in a preset historical time period, then input the historical behavior data into a click model for differential evolution calculation to obtain a test vector, which facilitates subsequent improvement of the accuracy of click rate estimation, then determine a target vector according to the test vector and the historical behavior data, so as to avoid falling into a local optimum in the prediction process, train the click model according to the target vector to obtain a click estimation model, which facilitates subsequent click rate estimation and improves the calculation precision of the click estimation model, finally, predict the click rate of the user's click behavior according to the click estimation model, and output demand information corresponding to the click behavior, so that the demand information that the user wants can be accurately predicted, and a display page corresponding to the demand information is recommended to the user, thereby increasing the browsing volume and access volume of the display page. BRIEF DESCRIPTION OF DRAWINGS

[0044] Figure 1 is a flowchart of the click rate estimation method based on a click model provided by the embodiment of the present application;

[0045] Figure 2 is a flowchart of step S102 in Figure 1

[0046] Figure 3 is a flowchart of step S202 in Figure 2

[0047] Figure 4 is a flowchart of step S203 in Figure 2

[0048] Figure 5 is a flowchart of step S103 in Figure 1

[0049] Figure 6 is a flowchart of step S104 in Figure 1

[0050] Figure 7 is a flowchart of the click rate estimation method based on a click model provided by another embodiment of the present application;

[0051] Figure 8 is a structural schematic diagram of the click rate estimation device based on a click model provided by the embodiment of the present application;

[0052] ​​​​​Figure 9 Fig. 1 is a schematic diagram of a hardware structure of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0053] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0054] It should be noted that although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a manner different from the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification and claims and the above-described drawings are used to distinguish similar objects, and do not necessarily describe a specific order or sequence.

[0055] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of the present application and are not intended to limit the present application.

[0056] First, several terms involved in the present application are analyzed:

[0057] Natural language processing (NLP): NLP uses computers to process, understand and use human languages (such as Chinese, English, etc.). NLP is a branch of artificial intelligence and is an interdisciplinary subject of computer science and linguistics, and is also commonly known as computational linguistics. Natural language processing includes syntax analysis, semantic analysis, discourse understanding, etc. Natural language processing is commonly used in machine translation, handwritten and printed character recognition, speech recognition and text-to-speech conversion, information intent recognition, information extraction and filtering, text classification and clustering, public opinion analysis and opinion mining, etc. It involves data mining, machine learning, knowledge acquisition, knowledge engineering, artificial intelligence research related to language processing, and language computing related linguistic research.

[0058] Differential Evolution (DE): Differential Evolution is a new evolutionary computation technique. DE algorithm is mainly used to solve global optimization problems of continuous variables. Its main working steps are basically the same as other evolutionary algorithms, mainly including mutation, crossover, and selection. The basic idea of the algorithm is to start from a randomly generated initial population, use the difference vector of two randomly selected individuals from the population as the random change source of the third individual, add the difference vector to the third individual according to certain rules to generate a mutant individual, which is called mutation. Then, the mutant individual and a pre-determined target individual are mixed to generate a trial individual, which is called crossover. If the fitness value of the trial individual is better than that of the target individual, the trial individual replaces the target individual in the next generation, otherwise the target individual is saved, which is called selection. In the evolution process of each generation, each individual vector is used as a target individual once, and the algorithm iteratively calculates to retain good individuals and eliminate poor individuals, guiding the search process to approach the global optimal solution.

[0059] Back Propagation Neural Network (BPNN): Back Propagation Neural Network is to increase several layers (one or more layers) of neurons between the input layer and the output layer. These neurons are called hidden units, which have no direct contact with the outside world, but their state changes can affect the relationship between input and output. Each layer can have several nodes. The basic back propagation algorithm includes two processes of forward propagation of signals and backward propagation of errors. That is, the error output is calculated in the direction from input to output, and the weight and threshold are adjusted in the direction from output to input. During forward propagation, input signals act on output nodes through hidden layers, and after nonlinear transformation, output signals are generated. If the actual output does not match the expected output, the error backward propagation process is entered. Error back propagation is to pass the output error through the hidden layer to the input layer layer by layer, and allocate the error to all units in each layer to obtain error signals from each layer as the basis for adjusting the weights of each unit. By adjusting the connection strength of input nodes and hidden nodes and the connection strength of hidden nodes and output nodes and the threshold, the error is reduced along the gradient direction. After repeated learning and training, the network parameters (weights and thresholds) corresponding to the minimum error are determined, and the training is stopped. At this time, the trained neural network can process input information similar to the sample and output error-minimized non-linearly converted information.

[0060] Click-Through-Rate (CTR): refers to the search engine input keywords after the search, and then according to the factors such as bidding, the relevant web pages are arranged in order, and then the user will select the website interested in clicking into; all the search times of a website are taken as the total number, and the proportion of the number of users clicking and entering the website to the total number is called click rate.

[0061] Error Function: In mathematics, the error function (also called the Gauss error function, Error Function or Gauss Error Function) is a non-basic function (i.e. not an elementary function) that has been widely used in probability theory, statistics, and partial differential equations and semiconductor physics. In statistics and probability theory, the Gaussian function is the density function of the normal distribution, and according to the central limit theorem, it is the finite probability distribution of the complex sum.

[0062] Activation Function: Activation function plays a very important role in artificial neural network model to learn and understand very complex and nonlinear functions. They introduce nonlinearity into our network. In neurons, inputs are weighted, summed, and then acted on a function, which is the activation function. The introduction of activation function is to increase the nonlinearity of neural network model. Without activation function, each layer is equivalent to matrix multiplication. Even if you stack several layers, it is nothing more than a matrix multiplication.

[0063] Based on this, the embodiments of the present application provide a click rate estimation method and device based on a click model, an electronic device and a storage medium, which can realize accurate estimation of click rate.

[0064] The click rate estimation method and device based on the click model provided by the embodiments of the present application, the electronic device and the storage medium are specifically explained through the following embodiments. First, the click rate estimation method based on the click model in the embodiments of the present application is described.

[0065] The embodiments of the present application can acquire and process related data based on artificial intelligence technology. Among them, artificial intelligence (Artificial Intelligence, AI) is to use digital computers or digital computer controlled machines to simulate, extend and expand human intelligence, perceive environment, acquire knowledge and use knowledge to obtain the best results. Theory, method, technology and application system.

[0066] The artificial intelligence basic technology generally includes technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics, etc. The artificial intelligence software technology mainly includes computer vision technology, robot technology, biometric identification technology, click rate estimation technology based on click model, natural language processing technology, and machine learning / deep learning, etc.

[0067] The click rate estimation method based on the click model provided by the embodiments of the present application relates to the technical field of data processing. The click rate estimation method based on the click model provided by the embodiments of the present application can be applied in a terminal, can also be applied in a server end, and can also be software running in the terminal or the server end. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, etc.; the server end can be configured as an independent physical server, can also be configured as a server cluster or a distributed system composed of multiple physical servers, can also be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, CDN, and big data and artificial intelligence platform, etc.; and the software can be an application implementing the click rate estimation method based on the click model, etc., but is not limited to the above forms.

[0068] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment, in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0069] It should be noted that in various specific embodiments of the present application, when relevant processing needs to be performed on data related to the identity or characteristics of the user, such as user information, user behavior data, user history data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of such data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain sensitive personal information of the user, the user's separate permission or separate consent will be obtained through a pop-up window or by jumping to a confirmation page, and after obtaining the separate permission or separate consent of the user, the necessary user-related data for enabling the embodiments of the present application to function normally will be obtained.

[0070] Figure 1 is an optional flowchart of a click rate estimation method based on a click model provided by the embodiments of the present application, Figure 1 The method in can include but is not limited to steps S101 to S106.

[0071] Step S101, obtaining historical behavior data of a user clicking a browser within a preset historical time length;

[0072] In step S101 of some embodiments, historical behavior data of a user clicking a browser within a preset historical time length is obtained, thereby facilitating subsequent training of a click model and improving the prediction accuracy of the model.

[0073] It should be noted that the historical behavior data of the user is historical exposure data in the user's browsing process, and the historical behavior data includes but is not limited to the user's age, gender, marital status, hobbies, historical click behavior, and click frequency, and the embodiments are not limited in particular.

[0074] It can be understood that the user's permission or consent will be obtained first in the process of obtaining the user's historical behavior data in the embodiments.

[0075] Step S102, inputting the historical behavior data into a click model for differential evolution to obtain a trial vector;

[0076] In step S102 of some embodiments, the historical behavior data is input into a click model for differential evolution calculation to obtain a trial vector, thereby realizing accelerated evolution of the historical behavior data.

[0077] It should be noted that the click model includes an input layer, a back propagation neural network, an output layer, neurons, and a plurality of nodes, and the embodiments are not limited in particular.

[0078] Step S103, determining a target vector according to the trial vector and the historical behavior data;

[0079] In step S103 of some embodiments, the target vector is determined according to the trial vector and the historical behavior data, so as to improve the convergence speed in the differential evolution process.

[0080] In step S104 of some embodiments, the click model is trained according to the target vector to obtain a click estimation model.

[0081] In step S104 of some embodiments, the click model is trained according to the target vector to obtain a click estimation model, so as to avoid falling into a local optimum in the training process and improve the estimation accuracy of the click behavior of the user.

[0082] In step S105, the click rate of the click behavior of the user is predicted according to the click estimation model, and the demand information corresponding to the click behavior is output.

[0083] In step S106, the display page corresponding to the demand information is recommended to the user.

[0084] The steps S101 to S106 shown in the embodiments of the present application obtain the historical behavior data of the user clicking the browser in a preset historical time period, and then input the historical behavior data into the click model for differential evolution calculation to obtain a trial vector, so as to facilitate subsequent improvement of the accuracy of the click rate estimation. Then, the target vector is determined according to the trial vector and the historical behavior data, so as to avoid falling into a local optimum in the prediction process. Then, the click model is trained according to the target vector to obtain a click estimation model, so as to facilitate subsequent estimation of the click rate and improve the calculation accuracy of the click estimation model. Finally, the click rate of the click behavior of the user is predicted according to the click estimation model, and the demand information corresponding to the click behavior is output. Thus, the demand information that the user wants can be accurately predicted, and the display page corresponding to the demand information is recommended to the user, so as to increase the browsing volume and access volume of the display page.

[0085] It should be noted that the embodiments can be applied in the click model test process. For example, before the click model is put into actual application scenarios (such as e-commerce scenarios, insurance scenarios, and medical scenarios), the model is tested. In the model test process, the model is tested by using the embodiments. The embodiments can also be applied to optimize the click model in the application after the click model is put into the actual application scenarios. For example, in the process of data prediction, the behavior data is input into the click model to optimize the click model to improve the prediction accuracy of the click model. The embodiments do not make specific limitations in this regard.

[0086] Please refer to Figure 2 In some embodiments, step S102 can include but is not limited to steps S201 to S203:

[0087] Step S201: Input historical behavior data into the click model for initialization calculation to obtain the initialization vector of historical behavior data and the optimal individual vector;

[0088] In step S201 of some embodiments, historical behavior data is input into the click model for initialization calculation to obtain the initialization vector of historical behavior data and the optimal individual vector. The calculation process of historical behavior data is shown in the following formula (1):

[0089] (1)

[0090] It should be noted that during the initial calculation of the historical behavior data input click model, NP random numbers are generated. D Individuals of a dimension: , xj,min ≤ xj,i,0 ≤ xj,max,j = 1,..., D, Each individual's The dimension is initialized using formula (1), where, and Representing the solution's first... The upper and lower bounds of a dimension, It represents a uniformly random number between 0 and 1.

[0091] Step S202: Perform mutation operation on historical behavior data based on the optimal individual vector and the pre-determined crossover vector to obtain the mutation vector;

[0092] In step S202 of some embodiments, the historical behavior data is mutated based on the optimal individual vector obtained in step S201 and the predetermined cross vector to obtain the mutated vector, wherein the predetermined cross vector is the result of the cross operation in the previous iteration.

[0093] Step S203: Perform a crossover operation on the mutation vector and the initialization vector to obtain the test vector.

[0094] In step S203 of some embodiments, the mutation vector and the initialization vector obtained in step S202 are cross-operated to obtain the test vector. The cross-operation of the mutation vector is shown in the following formula (3):

[0095] (3)

[0096] It should be noted that, For the test vector, For the mutation vector, To initialize the vector, is preset, wherein the preset crossover probability The greater the value of the preset crossover probability is, the greater the probability of crossover. In the embodiment, the preset crossover probability is 0.3.

[0097] Figure 3 In some embodiments, step S202 can include, but is not limited to, steps S301-S304.

[0098] In step S301, a current individual vector of historical behavior data is obtained.

[0099] In step S302, an optimal difference vector is determined according to the current individual vector and an optimal individual vector.

[0100] In step S303, an iteration difference vector is determined according to a crossover vector and the current individual vector.

[0101] In step S304, a mutation vector is obtained through a mutation operation based on a preset scaling factor, the optimal difference vector, the iteration difference vector, and the current individual vector.

[0102] The steps S301-S304 shown in the embodiments of the present application first obtain a current individual vector of historical behavior data , then determine an optimal difference vector according to the current individual vector and an optimal individual vector , then determine an iteration difference vector according to a crossover vector and the current individual vector, and finally obtain a mutation vector through a mutation operation based on a preset scaling factor , the optimal difference vector , the iteration difference vector , and the current individual vector .

[0103] It should be noted that the mutation operation of the optimal individual vector and the crossover vector is shown in the following formula (2):

[0104] (2)

[0105] wherein, is the optimal individual vector, is the difference vector between the optimal individual vector and the current individual, is the crossover vector, is the difference vector between the crossover vector and the individual, is the preset scaling factor.

[0106] It can be understood that the scaling factor The step size of the difference between population individuals determines the search of the optimal solution of the algorithm, and the value range is generally between [0.4, 0.95]. In the embodiment, the scaling factor is 0.6.

[0107] Referring to Figure 4 In some embodiments, step S203 can include but is not limited to step S401.

[0108] In step S401, the mutation vector and the initialization vector are subjected to a binomial crossover operation based on a preset crossover probability to obtain a test vector.

[0109] In step S401 of some embodiments, the mutation vector and the initialization vector are subjected to a binomial crossover operation based on a preset crossover probability to obtain a test vector, thereby guiding the individual vector to evolve in the optimal direction of the last iteration, improving the convergence speed, and avoiding the evolution of the individual vector in the direction of low click rate.

[0110] Referring to Figure 5 In some embodiments, step S103 can further include but is not limited to steps S501 to S505.

[0111] In step S501, a plurality of historical behavior vectors are determined according to historical behavior data.

[0112] In step S502, fitness calculation is performed on the historical behavior vectors to obtain historical fitness values of the historical behavior vectors.

[0113] In step S503, fitness calculation is performed on the test vector to obtain a test fitness value of the test vector.

[0114] In step S504, the historical fitness value and the test fitness value are compared to obtain a comparison result.

[0115] In step S505, a target vector is determined from the historical behavior vector and the test vector according to the comparison result.

[0116] ​​​The steps S501 to S505 shown in the embodiments of the present application first determine a plurality of historical behavior vectors corresponding to the historical behavior data according to the historical behavior data, then perform fitness calculation on the historical behavior vectors to obtain historical fitness values of the historical behavior vectors, and also need to perform fitness calculation on the test vectors to obtain test fitness values of the test vectors, finally compare the historical fitness values and the test fitness values to obtain a comparison result, determine whether to select the historical fitness values or the test fitness values according to the comparison result, and then determine the historical behavior vector corresponding to the historical fitness values or the test vector corresponding to the test fitness values as the target vector, so as to complete the selection operation of the historical behavior vector and the test vector.

[0117] It should be noted that in the process of comparing the historical fitness values and the test fitness values, the vector with the smaller fitness value between the two is retained to enter the next generation and become the parent of the next generation, that is, the target vector in the present embodiment, so as to avoid falling into local optimum.

[0118] It can be understood that the selection operation of the historical behavior vector and the test vector is shown in the following formula (4):

[0119] (4)

[0120] wherein, is the test vector, and the test vector is represented as , is the historical behavior vector, and the historical behavior vector can be represented as , is the target vector determined according to the test vector and the historical behavior vector.

[0121] Please refer to Figure 6 In some embodiments, step S104 includes but is not limited to steps S601 to S603:

[0122] It should be noted that the click model includes a back propagation neural network.

[0123] Step S601, gradient descent calculation is performed on the back propagation neural network to obtain network weights and network thresholds of the back propagation neural network;

[0124] In step S601 of some embodiments, gradient descent calculation is performed on the back propagation neural network to obtain network weights and network thresholds of the back propagation neural network, which facilitates subsequent training of the network weights and the network thresholds, so as to realize optimization of the back propagation neural network and improve the convergence speed and prediction accuracy of the back propagation neural network.

[0125] It should be noted that the back propagation neural network includes a plurality of structure parameters, for example, the number of input layer neurons, the number of hidden layer neurons, the number of output layer neurons, an activation function and the like, wherein the network weight and the network threshold are calculated by descending according to the preset error function of the back propagation neural network as a gradient function.

[0126] In step S602, the network weight and the network threshold are trained according to the target vector to obtain the fitness function.

[0127] In step S602 of some embodiments, the network weight and the network threshold are trained according to the target vector to obtain the fitness function, so as to realize the optimization of the back propagation neural network and improve the convergence speed and the estimation accuracy of the back propagation neural network.

[0128] In step S603, the click model is trained according to the fitness function to obtain the click estimation model.

[0129] In step S603 of some embodiments, the click model is trained according to the fitness function to obtain the click estimation model, and the network weight and the network threshold are optimized by the differential evolution algorithm, so as to improve the prediction accuracy of the click estimation model.

[0130] Figure 7 is a flowchart of a click rate estimation method based on a click model provided by another embodiment of the present application, Figure 7 The method in the above formula (1) can include but is not limited to including step S701.

[0131] In step S701, the initialization vector is updated every preset optimization times.

[0132] In step S701 of some embodiments, in order to avoid falling into local optimum, the initialization vector needs to be updated every preset optimization times, so as to realize the replacement of the individual vector and increase the diversity of the initialization vector.

[0133] It should be noted that the preset optimization times can be 5 times, 8 times, 10 times and the like, and the preset optimization times in the present embodiment is 10 times, that is, after 10 differential evolutions, 10%NP individual vectors are randomly generated by using the formula (1) and 10% of the existing vectors are replaced, wherein the replacement process can be directly generating 10%NP new individual vectors, or randomly extracting 10% of the individual vectors from the original individual vectors and replacing them with the new individual vectors obtained in the last step, and the present embodiment does not make specific limitation.

[0134] It is worth noting that the initialization calculation of the historical behavior data, the mutation operation of the historical behavior data, the crossover operation of the initialization vector, the operation of determining the target vector and the updating operation of the initialization vector need to be repeatedly performed in the embodiment, and the above steps are repeatedly performed until the condition for stopping the algorithm is met, so that the accurate prediction of the user click rate is realized, and the local optimum is avoided.

[0135] It should be noted that the condition for stopping the algorithm can be that the fitness value no longer changes or the number of iterations reaches a preset iteration value, and the embodiment does not make specific limitations.

[0136] Please refer to Figure 8 The embodiment of the application also provides a click rate prediction device based on a click model, which can implement the click rate prediction method based on the click model. The device comprises:

[0137] The data acquisition module 801 is configured to acquire historical behavior data of a user clicking a browser within a preset historical time period.

[0138] The differential evolution module 802 is configured to input the historical behavior data into the click model to perform differential evolution calculation to obtain a trial vector.

[0139] The vector determination module 803 is configured to determine a target vector according to the trial vector and the historical behavior data.

[0140] The model training module 804 is configured to train the click model according to the target vector to obtain a click prediction model.

[0141] The data prediction module 805 is configured to predict a click rate of a user's click behavior according to the click prediction model and output demand information corresponding to the click behavior.

[0142] The page display module 806 is configured to recommend a display page corresponding to the demand information to the user.

[0143] The specific implementation of the click rate prediction device based on the click model is basically the same as that of the specific embodiment of the click rate prediction method based on the click model, and will not be repeated here.

[0144] The embodiment of the application also provides an electronic device, which comprises a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for realizing the connection and communication between the processor and the memory. The program is executed by the processor to implement the click rate prediction method based on the click model. The electronic device can be any intelligent terminal, such as a tablet computer or a vehicle-mounted computer.

[0145] Please refer to Figure 9 , Figure 9The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0146] The processor 901 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits, and is configured to execute related programs to implement the technical solutions provided by the embodiments of the present application.

[0147] The memory 902 can be implemented by a ROM (Read Only Memory), a static storage device, a dynamic storage device, or a RAM (Random Access Memory), and the like. The memory 902 can store an operating system and other application programs, and when the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are stored in the memory 902 and are called and executed by the processor 901 to implement the click rate estimation method based on the click model.

[0148] The input / output interface 903 is configured to realize information input and output.

[0149] The communication interface 904 is configured to realize the communication interaction between the device and other devices, and can realize the communication through a wired manner (for example, a USB, a network cable, or the like) or a wireless manner (for example, a mobile network, WIFI, Bluetooth, or the like).

[0150] The bus 905 is configured to transmit information between the components (for example, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904) of the device.

[0151] The processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are connected to each other through the bus 905 to realize the communication connection between the components in the device.

[0152] The embodiments of the present application further provide a computer readable storage medium, which stores a computer program. The computer program is executed by a processor to implement the click rate estimation method based on the click model.

[0153] The memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include a high-speed random access memory and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory that is remotely arranged relative to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0154] The click rate estimation method and device based on a click model, the electronic device and the storage medium provided by the embodiments of the present application first acquire historical behavior data of a user clicking a browser within a preset historical time period, then input the historical behavior data into a click model for differential evolution calculation to obtain a trial vector, which facilitates subsequent improvement of the accuracy of click rate estimation, then determine a target vector according to the trial vector and the historical behavior data, so as to avoid falling into a local optimum in the prediction process, train the click model according to the target vector to obtain a click estimation model, which facilitates subsequent click rate estimation and improves the calculation precision of the click estimation model, finally, predict the click rate of the user's click behavior according to the click estimation model, and output demand information corresponding to the click behavior, so that the demand information that the user wants can be accurately predicted, and a display page corresponding to the demand information is recommended to the user, thereby increasing the browsing volume and access volume of the display page.

[0155] The embodiments described in the embodiments of the present application are used to more clearly illustrate the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0156] Those skilled in the art can understand that, Figures 1-7 The technical solutions shown in the above description do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0157] The device embodiments described above are only schematic, and the units described as separate components can or can not be physically separate, that is, can be located in one place, or can be distributed on multiple network units. Part or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0158] Those skilled in the art can understand that all or some of the steps in the method disclosed above, the function modules / units in the system and the device can be implemented as software, firmware, hardware or appropriate combination thereof.

[0159] The terms "first", "second", "third", "fourth" etc. (if any) in the description of the application and in the claims that follow are used for distinguishing between similar elements and not necessarily for describing a sequential or chronological order. It is to be understood that the use of these terms herein is to be construed to cover the embodiments of the application whether or not the embodiments are described using the same term. Furthermore, the terms "comprise", "comprising", "include", "including", and "has", "having" and variants thereof are to be construed in a non-exclusive manner when used in this description and in the claims that follow. For example, when used in the context of a process, method, system, product or apparatus, the term "comprising" means that the process, method, system, product or apparatus includes the recited steps or units, but can also include additional steps or units not specifically recited.

[0160] It should be understood that, in the present application, "at least one" means one or more, and "multiple" means two or more. "And / or" is used to describe the relationship between associated objects, which means that there can be three relationships, for example, "A and / or B" can mean that there are three cases: only A, only B, and A and B at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or similar expressions means any combination of these items, including any combination of single or multiple items. For example, at least one of a, b or c can mean a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be singular or plural.

[0161] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above-mentioned units is only a logical function division, and actual implementation can have another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be omitted or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed objects can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0162] The units described as separate components above can or can not be physically separate, and the components shown as units can or can not be physical units, that is, can be located in one place, or can be distributed to multiple network units. Part or all of the units can be selected according to actual needs to achieve the purpose of the embodiment scheme.

[0163] In addition, each functional unit in each embodiment of the present application can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0164] If the integrated unit is realized in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on this understanding, the technical scheme of the present application or the part that contributes to the prior art or the whole or part of the technical scheme can be embodied in the form of a software product. The computer software product is stored in a storage medium, including multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the method of each embodiment of the present application. The foregoing storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0165] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, but this does not limit the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A click rate estimation method based on a click model, characterized in that, The method comprises: acquiring historical behavior data of a user clicking a browser in a preset historical time length; inputting the historical behavior data into the click model for differential evolution to obtain a trial vector; determining a target vector according to the trial vector and the historical behavior data; training the click model according to the target vector to obtain a click estimation model; predicting a click rate of a click behavior of the user according to the click estimation model, and outputting demand information corresponding to the click behavior; recommending a display page corresponding to the demand information to the user; wherein the trial vector is obtained by performing cross operation on a mutation vector and an initialized vector of the historical behavior data, and the mutation vector is obtained by the following steps: acquiring a current individual vector of the historical behavior data; determining an optimal difference vector according to the current individual vector and an optimal individual vector; determining an iteration difference vector according to a predetermined cross vector and the current individual vector; performing mutation operation based on a preset scaling factor, the optimal difference vector, the iteration difference vector and the current individual vector to obtain the mutation vector.

2. The click rate estimation method based on a click model according to claim 1, characterized in that, The inputting the historical behavior data into the click model for differential evolution calculation to obtain a trial vector comprises: inputting the historical behavior data into the click model for initialization calculation to obtain an initialized vector of the historical behavior data and an optimal individual vector; performing mutation operation on the historical behavior data according to the optimal individual vector and a cross vector to obtain a mutation vector; performing cross operation on the mutation vector and the initialized vector to obtain the trial vector.

3. The click rate estimation method based on a click model according to claim 2, characterized in that, The performing cross operation on the mutation vector and the initialized vector to obtain the trial vector comprises: performing binomial cross operation on the mutation vector and the initialized vector based on a preset cross probability to obtain the trial vector.

4. The click model based click rate estimation method of claim 1, wherein, The determining a target vector according to the trial vector and the historical behavior data comprises: determining a plurality of historical behavior vectors according to the historical behavior data; performing fitness calculation on the historical behavior vectors to obtain historical fitness values of the historical behavior vectors; performing fitness calculation on the trial vector to obtain a trial fitness value of the trial vector; comparing the historical fitness values and the trial fitness value to obtain a comparison result; determining the target vector from the historical behavior vectors and the trial vector according to the comparison result.

5. The click model based click rate estimation method of claim 1, wherein, The click model comprises a back propagation neural network; the training the click model according to the target vector to obtain a click estimation model comprises: performing gradient descent calculation on the back propagation neural network to obtain network weights and network thresholds of the back propagation neural network; training the network weights and the network thresholds according to the target vector to obtain a fitness function; training the click model according to the fitness function to obtain the click estimation model.

6. The click model based click rate estimation method of claim 2, wherein, After the training the click model according to the target vector to obtain a click estimation model, the method further comprises: updating the initialized vector every preset optimization times. 7.A click rate estimation device based on a click model, characterized in that, The device comprises: a data acquisition module configured to acquire historical behavior data of a user clicking a browser in a preset historical time period; a differential evolution module configured to input the historical behavior data into the click model to perform differential evolution calculation to obtain a trial vector; a vector determination module configured to determine a target vector according to the trial vector and the historical behavior data; a model training module configured to train the click model according to the target vector to obtain a click estimation model; a data prediction module configured to perform click rate prediction on a click behavior of the user according to the click estimation model and output demand information corresponding to the click behavior; a page display module configured to recommend a display page corresponding to the demand information to the user; wherein the trial vector is obtained by performing cross operation on a mutation vector and an initial vector of the historical behavior data, and the mutation vector is obtained by the following steps: acquiring a current individual vector of the historical behavior data; determining an optimal difference vector according to the current individual vector and an optimal individual vector; determining an iteration difference vector according to a predetermined cross vector and the current individual vector; performing mutation operation based on a preset scaling factor, the optimal difference vector, the iteration difference vector and the current individual vector to obtain the mutation vector.

8. An electronic device, comprising: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the click rate estimation method based on the click model according to any one of claims 1 to 6 when executing the computer program.

9. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The computer program is executed by the processor to implement the click rate estimation method based on the click model according to any one of claims 1 to 6.

Images