WO2019120019A1

WO2019120019A1 - User gender prediction method and apparatus, storage medium and electronic device

Info

Publication number: WO2019120019A1
Application number: PCT/CN2018/116481
Authority: WO
Inventors: 陈岩; 刘耀勇
Original assignee: Oppo广东移动通信有限公司
Priority date: 2017-12-20
Filing date: 2018-11-20
Publication date: 2019-06-27
Also published as: CN109948633A

Abstract

Disclosed is a user gender prediction method, comprising: collecting multi-dimensional feature information of a user whose gender is already known using an electronic device so as to construct a sample set; training a BP neural network model by using the sample set to obtain a prediction model; obtaining multi-dimensional feature information of a user whose gender is unknown using the electronic device as a prediction sample; and generating a prediction probability according to the prediction sample and the prediction model. The present application further provides a user gender prediction apparatus, a storage medium, and an electronic device.

Description

User gender prediction method, device, storage medium and electronic device

This application claims the priority of the Chinese Patent Application filed on Dec. 20, 2017, the Chinese Patent Office, Application No. 201711387919.0, entitled "User Gender Prediction Method, Apparatus, Storage Media, and Electronic Equipment", the entire contents of which are incorporated by reference. In this application.

Technical field

The present application belongs to the field of communications technologies, and in particular, to a user gender prediction method, apparatus, storage medium, and electronic device.

Background technique

With the development of electronic device technology, electronic devices have begun to gradually become a platform for running general-purpose software from a simple provision of a call device. The platform is no longer to provide call management as the main purpose, but to provide an operating environment including call management, game entertainment, office notes, mobile payment and other application software, with a large number of popularization, has penetrated into people's Every aspect of life and work.

User portraits are a very popular research direction in recent years. For example, on a smart phone, if there is a way to accurately determine the user's gender from the user's behavioral habits, it is very meaningful to optimize the mobile phone in all aspects.

Summary of the invention

The present application provides a user gender prediction method, apparatus, storage medium, and electronic device capable of predicting a user's gender based on a BP neural network.

In a first aspect, an embodiment of the present application provides a user gender prediction method, including:

Collecting multi-dimensional feature information of a known gender user using the electronic device as a sample, and constructing a sample set of gender prediction;

Using the sample set to train the BP neural network model to obtain a trained prediction model;

Obtaining multi-dimensional feature information of an electronic device that is used by an unknown gender user as a prediction sample;

A prediction probability is generated according to the predicted sample and the trained prediction model, the predicted probability including: a first probability that the current user is a male, and a second probability that the current user is a female.

In a second aspect, the embodiment of the present application provides a user gender prediction apparatus, including:

An acquisition module is configured to collect multi-dimensional feature information of a known gender user using the electronic device as a sample, and construct a sample set of gender prediction;

a training module, configured to train the BP neural network model by using the sample set to obtain a trained prediction model;

An obtaining module, configured to obtain multi-dimensional feature information of an electronic device used by an unknown gender user as a prediction sample;

And a generating module, configured to generate a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability includes: a first probability that the current user is a male, and a second probability that the current user is a female.

In a third aspect, an embodiment of the present application provides a storage medium on which a computer program is stored, and when the computer program runs on a computer, causes the computer to execute the user gender prediction method described above.

In a fourth aspect, an embodiment of the present application provides an electronic device, including a processor and a memory, where the memory stores a plurality of instructions, and the processor loads the instructions in the memory to perform the following steps:

DRAWINGS

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings used in the description of the embodiments will be briefly described below. Obviously, the drawings in the following description are only some embodiments of the present application, and those skilled in the art can obtain other drawings according to the drawings without any creative work.

FIG. 1 is a schematic diagram of a system for a user gender prediction apparatus according to an embodiment of the present application.

FIG. 2 is a schematic diagram of an application scenario of a user gender prediction apparatus according to an embodiment of the present application.

FIG. 3 is a schematic flowchart of a user gender prediction method according to an embodiment of the present application.

FIG. 4 is another schematic flowchart of a user gender prediction method according to an embodiment of the present application.

FIG. 5 is a schematic diagram of another application scenario of a user gender prediction apparatus according to an embodiment of the present application.

FIG. 6 is a schematic structural diagram of a user gender prediction apparatus according to an embodiment of the present application.

FIG. 7 is another schematic structural diagram of a user gender prediction apparatus according to an embodiment of the present application.

FIG. 8 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

FIG. 9 is another schematic structural diagram of an electronic device according to an embodiment of the present application.

Detailed ways

Referring to the drawings, wherein like reference numerals represent the same components, the principles of the present application are illustrated by the implementation in a suitable computing environment. The following description is based on the specific embodiments of the present invention as illustrated, and should not be construed as limiting the specific embodiments that are not described herein.

In the following description, specific embodiments of the present application will be described with reference to the steps and symbols executed by one or more computers, unless otherwise stated. Thus, these steps and operations will be referred to several times by a computer, and the computer execution referred to herein includes the operation of a computer processing unit that represents an electronic signal in data in a structured version. This operation converts the data or maintains it at a location in the computer's memory system, which can be reconfigured or otherwise altered in a manner well known to those skilled in the art. The data structure maintained by the data is the physical location of the memory, which has specific characteristics defined by the data format. However, the principles of the present application are described in the above text, which is not intended to be a limitation, and those skilled in the art will appreciate that the various steps and operations described below can also be implemented in hardware.

The terms "first," "second," and "third," etc. in this application are used to distinguish different objects, and are not intended to describe a particular order. Furthermore, the terms "comprises" and "comprising" and "comprising" are intended to cover a non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or modules is not limited to the listed steps or modules, but some embodiments also include steps or modules not listed, or some embodiments Other steps or modules inherent to these processes, methods, products or devices are also included.

In today's streaming electronic device systems, users are allowed to register binding devices and user accounts. But not every user is willing to provide gender information. Therefore, this method cannot solve the problem of a large part of the user's portrait. The electronic device may be a smart phone, a tablet computer, a desktop computer, a notebook computer, or a handheld computer.

Please refer to FIG. 1. FIG. 1 is a schematic diagram of a system for predicting a user gender according to an embodiment of the present application. The user gender prediction device is mainly used for collecting multi-dimensional feature information of a known gender user using the electronic device as a sample, constructing a sample set of gender prediction, and training the BP neural network model by using the sample set to obtain a trained prediction model. Obtaining multi-dimensional feature information of the unknown gender user using the electronic device as a prediction sample, and generating a prediction probability according to the prediction sample and the trained prediction model, the prediction probability includes: a first probability that the current user is a male, and a second probability that the current user is a female Probability.

Specifically, please refer to FIG. 2 , which is a schematic diagram of an application scenario of a user gender prediction apparatus according to an embodiment of the present application. For example, when receiving the prediction request, the user gender prediction device collects multi-dimensional feature information of the electronic device that the unknown gender user uses. The multi-dimensional feature information may be used to acquire historical usage information in the electronic device and collect multi-dimensional feature information from the historical usage information. The multi-dimensional feature information is input to the prediction model, and the current user is predicted by the prediction model to obtain the first probability. The prediction model is used to predict the current user's female probability, and the second probability is obtained, and finally according to the first probability sum. The second probability is entered into the final predicted result.

The embodiment of the present application provides a user gender prediction method, and the execution subject of the user gender prediction method may be the user gender prediction device provided by the embodiment of the present application, or an electronic device integrated with the user gender prediction device, wherein the user gender prediction The device can be implemented in hardware or software.

The embodiment of the present application will be described from the perspective of a user gender prediction device, which may be specifically integrated in an electronic device. The user gender prediction method includes: collecting multi-dimensional feature information of a known gender user using the electronic device as a sample, constructing a sample set of gender prediction, and training the BP neural network model by using the sample set to obtain a trained prediction model and acquiring an unknown. The gender user uses the multi-dimensional feature information of the electronic device as a prediction sample, and generates a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability includes: the first probability that the current user is a male, and the current user is a female Two probability.

In an embodiment, the step of training the BP neural network model by using the sample set comprises:

Normalizing the sample parameters in the sample set;

The BP neural network model is trained according to the normalized sample parameters.

In an embodiment, the step of training the BP neural network model according to the normalized sample parameters includes:

Entering the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results;

Obtaining a loss value according to the two prediction results and a probability corresponding thereto;

Training is performed based on the loss value to generate a target model parameter.

In an embodiment, the step of training according to the loss value includes:

Training is performed using the gradient descent method based on the loss value.

In an embodiment, after generating the prediction probability according to the predicted sample and the trained prediction model, the method further includes:

Comparing a first probability that the current user is a male with a second probability that the current user is a female, and obtaining a comparison result;

Based on the comparison result, the final prediction result is output.

In an embodiment, the step of outputting the final prediction result according to the comparison result includes:

When the first probability is greater than the second probability, outputting a first prediction result that the current user is a male;

When the first probability is not greater than the second probability, outputting a second prediction result that the current user is a female.

In one embodiment, the step of inputting the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results includes:

Calculating a probability based on a first preset formula to obtain two prediction results, wherein the first preset formula is:

Where Z _K is the intermediate value and C is the number of categories of the predicted result.

For the jth intermediate value, a two-dimensional output layer is obtained based on the first preset formula, respectively representing the probability that the sample is male and female.

In an embodiment, the step of obtaining a loss value according to the two prediction results and the probability corresponding thereto includes:

And obtaining, according to the second preset formula, a loss value according to the two prediction results and a probability corresponding thereto, wherein the second preset formula is:

Where C is the number of categories of prediction results, y _k is the true value, and E is the average.

Please refer to FIG. 3. FIG. 3 is a schematic flowchart diagram of a user gender prediction method according to an embodiment of the present application. The user gender prediction method provided by the embodiment of the present application is applied to an electronic device, and the specific process may be as follows:

Step 101: Collect multi-dimensional feature information of a known gender user using the electronic device as a sample, and construct a sample set of gender prediction.

It is known that the multi-dimensional feature information of the electronic user using the electronic device has a dimension of a certain length, and the parameters in each dimension correspond to one feature information of the characterization application, that is, the multi-dimensional feature information is composed of a plurality of feature information. The plurality of feature information may include feature information of the user in using different types of applications, such as the number and duration of browsing the male-type goods (such as men's clothing) in the shopping application, and the user browsing the female-oriented products in the shopping application (such as The number and duration of cosmetics and women's wear; the length of time users read a male novel in a reading application, the length of time a user reads a female novel, the length of time a user reads a sports news, the length of time a user reads a constellation news; the user uses a different type The usage information in the application, such as the number of times the user uses the front camera self-timer, the number of times the user uses the beauty software, and the number and duration of the user playing different types of games.

The sample set of gender predictions may include multiple training samples collected during the historical time period. The historical time period can be, for example, the past 7 days, 10 days, and the like. It can be understood that the known gender users collected at one time use the multi-dimensional feature data of the electronic device to form a sample set. The feature set for each sample is recorded with a set of real numbers. The sample values need to be normalized during sampling, such as normalization to a number between 0 and 1. Thus each user is a sample, and N users make up N samples, each of which is characterized by (x ₁ , x ₂ , . . . , x _n ).

After constituting the sample set, each sample in the sample set can be marked to obtain a sample label of each sample. Since the implementation is to predict the gender of the current user, the labeled sample label is the registered account of the user. The actual gender information provided, including gender males and gender females, can be represented by a code of 0 or 1. In this case, the sample categories can include gender males and gender females. Specifically, the user may mark the historical usage habits of the application in the electronic device, for example, within a length of one week, the user browses the male-type goods (such as men's clothing) in the shopping application for 20 times, and the accumulated duration is one hour. It can be marked as a gender male; for example, within a period of one week, the user can read a female novel for 8 hours, which can be marked as a gender female. Specifically, the value "1" can be used to indicate "gender male". ", use the value "0" to indicate "gender female", and vice versa.

Wherein, the sample set of the above gender prediction is used to train the BP neural network model, for example, the sample set may include the following features:

The number and duration of browsing of male-type items (such as men's clothing) in the shopping app;

The number and duration of browsing of female-type items (such as cosmetics and women's clothing) in the shopping application;

The length of time users read a male-like novel;

The length of time users read a female novel;

The length of time users read sports news;

The length of time users read the news of the constellation;

The number of times the user used the front camera to take a self-portrait;

The number of times the user uses the beauty software;

The number and duration of users playing different types of games.

In step 102, the BP neural network model is trained by using the sample set to obtain the trained prediction model.

Among them, BP neural network model is a classification model in machine learning. Its basic idea is gradient descent method, which uses gradient search technology to minimize the error mean square error between the actual output value of the network and the expected output value.

The basic BP algorithm includes two processes of forward propagation of signals and back propagation of errors. That is, the calculation of the error output is performed in the direction from the input to the output, and the adjustment weight and the threshold are performed from the output to the input. In the case of forward propagation, the input signal acts on the output node through the hidden layer and undergoes a nonlinear transformation to produce an output signal. If the actual output does not match the expected output, the error propagates back into the error propagation process. Error back propagation is to pass the output error back to the input layer through the hidden layer, and distribute the error to all the units in each layer, so as to adjust the error value of each unit as the basis for adjusting the weight of each unit. By adjusting the connection strength between the input node and the hidden layer node and the connection strength and threshold of the hidden layer node and the output node, the error is decreased along the gradient direction, and after repeated learning training, the network parameters (weights and thresholds corresponding to the minimum error) are determined. ), the training will stop. At this time, the trained neural network can process the non-linear conversion information with the smallest output error for the input information of similar samples.

In the embodiment of the present application, the BP neural network model can be trained by using the sample set to obtain the trained prediction model. Among them, the training model of the BP neural network model is used to solve the model parameters in the BP neural network model.

The network structure of the BP neural network model in the embodiment of the present application includes three layers, which are an input layer, a hidden layer, and an output layer, respectively, wherein the input layer inputs the feature information in step 101, and of course the feature information is normalized. After the feature information, for example, the above 9 feature information is introduced into 9 input nodes, and the hidden layer may include three layers, which may be 10 nodes, 5 nodes, 2 nodes respectively, and finally obtain 2D output through the SoftMax function. The layers represent the probability that the user is male and female.

For all samples collected, they were sent to the network for training in batches using the mini batch gradient descent method. The above mini batch gradient descent method is an optimization algorithm between the fastest gradient descent method and the stochastic gradient descent method, and each time a certain amount of training samples are selected for iteration. The cross-entropy (crossentropy) is used as the loss function to update the network weights in reverse propagation. After the loss value is less than the preset threshold or the number of iterations reaches the set number of training iterations, the training can be ended.

Wherein, the loss function is used to estimate the degree of inconsistency between the predicted value f(x) of the model and the true value Y. It is a non-negative real-valued function, usually using L(Y, f(x)), or L(w) indicates that the smaller the loss function, the better the robustness of the model. The loss function is the core part of the empirical risk function and an important part of the structural risk function.

Step 103: Acquire the current multi-dimensional feature information of the electronic device of the unknown gender user and use it as a prediction sample.

For example, the user characteristic value mentioned in the previous week is counted, the user feature vector is obtained, and the normalized feature value is input into the network to perform a forward calculation, and the output of the network is the probability that the user gender is male or female. The following characteristics can be included in the forecast sample:

The length of time users read a male-like novel;

The length of time users read a female novel;

The length of time users read sports news;

The length of time users read the news of the constellation;

The number of times the user used the front camera to take a self-portrait;

The number of times the user uses the beauty software;

The number and duration of users playing different types of games.

Step 104: Generate a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability includes: a first probability that the current user is a male, and a second probability that the current user is a female.

According to the prediction set machine and its corresponding trained BP neural network model, a corresponding probability is output, and the BP neural network model outputs a first probability that the current user is a male and a second probability that the current user is a female.

In an embodiment, in order to simplify the application prediction operation and increase the prediction speed, one probability may be selected for two probabilities in the prediction probability, and then the gender of the current user is predicted based on the selected probability. For example, after generating the prediction probability according to the prediction sample and the trained prediction model, the method may further include:

Comparing the first probability that the current user is a male with the second probability that the current user is a female, and obtaining a comparison result;

According to the comparison result, the final prediction result is output.

In an embodiment, according to the comparison result, the step of outputting the final prediction result may include:

When the first probability is not greater than the second probability, the second predicted result that the current user is a female is output.

For example, for a certain prediction probability, if Y=1 indicates that the current user is male, Y=0 indicates that the current user is female, and if P(Y=1|x) is greater than P(Y=0|x), then the output is The current user is the first predicted result of the male; assuming P(Y=1|x) is not greater than P(Y=0|x), at this time, the second predicted result that the current user is a female is output.

It should be noted that the training process of the predictive model can be completed on the server side or on the electronic device side. When the training process and the actual prediction process of the predictive model are completed on the server side, when the trained predictive model is needed, the multi-dimensional feature information of the unknown gender user using the electronic device can be input to the server, and after the actual prediction of the server is completed, the prediction will be performed. The result is sent to the electronic device end, and finally the electronic device outputs the predicted result.

When the training process and the actual prediction process of the predictive model are completed on the electronic device end, when the trained predictive model is needed, the multi-dimensional feature information can be input to the electronic device, and after the actual prediction of the electronic device is completed, the electronic device outputs the predicted result. .

As can be seen from the above, the embodiment of the present application collects multi-dimensional feature information of a known gender user using the electronic device as a sample, constructs a sample set of gender prediction, and uses the sample set to train the BP neural network model to obtain a trained prediction model and obtain The unknown gender user uses the multi-dimensional feature information of the electronic device as a prediction sample, and generates a prediction probability according to the predicted sample and the trained prediction model, the predicted probability includes: the first probability that the current user is a male, and the current user is a female Second probability. The application can predict the gender of the current user based on the BP neural network model, improve the accuracy of the gender prediction, and facilitate the electronic device to complete the accurate portrait of the user.

The cleaning method of the present application will be further described below based on the method described in the above embodiments. Referring to FIG. 4, FIG. 4 is another schematic flowchart of a user gender prediction method according to an embodiment of the present application, where the user gender prediction method includes:

Step 201: Collect multi-dimensional feature information of a known gender user using the electronic device as a sample, and construct a sample set of gender prediction.

Step S202, normalizing the sample parameters in the sample set.

The feature set for each sample is recorded with a set of real numbers. The sample values need to be normalized during sampling, such as normalization to a number between 0 and 1. Thus each user is a sample, and N users make up N samples, each of which is characterized by (x ₁ , x ₂ , . . . , x _n ).

Step S203, input the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results.

The network structure of the BP neural network model in the embodiment of the present application includes three layers, which are an input layer, a hidden layer, and an output layer, respectively, wherein the input layer inputs the feature information in step 101, and of course the feature information is normalized. After the feature information, for example, the above 9 feature information is introduced into 9 input nodes, and the hidden layer may include three layers, which may be 10 nodes, 5 nodes, 2 nodes respectively, and finally obtain 2D output through the SoftMax function. The layers represent the probability that the user is male and female. For all samples collected, they were sent to the network for training in batches using the mini batch gradient descent method.

In this embodiment, the probability of obtaining two prediction results is calculated based on the first preset formula, wherein the first preset formula is:

Step S204, the loss value is obtained according to the two prediction results and the probability corresponding thereto.

In this embodiment, the loss value is obtained according to the two prediction results and the probability corresponding thereto according to the second preset formula, wherein the second preset formula is:

In step S205, training is performed according to the loss value to generate a target model parameter.

Among them, the gradient descent method is used for training according to the loss value. In this embodiment, all the samples collected are sent to the network for training in a batch manner using a mini batch gradient descent method. The above mini batch gradient descent method is an optimization algorithm between the fastest gradient descent method and the stochastic gradient descent method, and each time a certain amount of training samples are selected for iteration. The cross-entropy (crossentropy) is used as the loss function to update the network weights in reverse propagation. After the loss value is less than the preset threshold or the number of iterations reaches the set number of training iterations, the training can be ended.

Step S206: Acquire multi-dimensional feature information of the electronic device of the unknown gender user and use it as a prediction sample.

The length of time users read a male-like novel;

The length of time users read a female novel;

The length of time users read sports news;

The length of time users read the news of the constellation;

The number of times the user used the front camera to take a self-portrait;

The number of times the user uses the beauty software;

The number and duration of users playing different types of games.

Step S207, generating a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability comprises: a first probability that the current user is a male, and a second probability that the current user is a female.

In step S208, the first probability that the current user is a male is compared with the second probability that the current user is a female, and a comparison result is obtained.

Step S209, according to the comparison result, output the final prediction result.

In this embodiment, according to the comparison result, the step of outputting the final prediction result may include:

As can be seen from the above, the embodiment of the present application collects the multi-dimensional feature information of the known gender user using the electronic device as a sample, constructs a sample set of the gender prediction, normalizes the sample parameters in the sample set, and normalizes the sample parameters. The BP neural network model is input to obtain the probability of two prediction results, the loss value is obtained according to the two prediction results and the probability corresponding thereto, the training is performed according to the loss value, the target model parameter is generated, and the multi-dimensional use of the electronic device by the unknown gender user is obtained. The feature information is used as a prediction sample, and the prediction probability is generated according to the prediction sample and the trained prediction model. The prediction probability includes: a first probability that the current user is a male, and a second probability that the current user is a female, and the current user is a male. A probability is compared with a second probability that the current user is a woman, a comparison result is obtained, and a final prediction result is output according to the comparison result. The application can predict the gender of the current user based on the BP neural network model, improve the accuracy of the gender prediction, and facilitate the electronic device to complete the accurate portrait of the user.

Referring to FIG. 5, FIG. 5 is a schematic diagram of another application scenario of a user gender prediction apparatus according to an embodiment of the present application. When the training process of the predictive model is completed on the server side, and the actual prediction process of the predictive model is completed on the electronic device side, when the optimized predictive model needs to be used, the current multidimensional feature information of the electronic device can be input to the electronic device, and the electronic device actually After the prediction is completed, the electronic device outputs the predicted result based on the predicted result. Optionally, the trained predictive model file (model file) can be transplanted to the smart device. If it is necessary to determine the gender of the current user, update the current sample set, input the trained predictive model file (model file), and calculate You can get the predicted value.

In some embodiments, before the step of acquiring the current multi-dimensional feature information of the electronic device, the method may further include:

The preset time is obtained. If the current system time reaches the preset time, the current multi-dimensional feature information of the electronic device is obtained. The preset time can be a time point in the day, such as 9 am, or several time points in the day, such as 9 am, 6 pm, and the like. It can also be one or several time points in multiple days. Then, the prediction result is generated according to the prediction model and the optimization parameter, and the application is controlled according to the prediction result.

All of the above technical solutions may be combined to form an optional embodiment of the present application, and will not be further described herein.

Please refer to FIG. 6. FIG. 6 is a schematic structural diagram of a user gender prediction apparatus according to an embodiment of the present application. The user gender prediction device 300 is applied to an electronic device, and the user gender prediction device 300 includes an acquisition module 301, a training module 302, an acquisition module 303, and a generation module 304.

The collecting module 301 is configured to collect multi-dimensional feature information of the known gender user using the electronic device as a sample, and construct a sample set of the gender prediction.

Specifically, the sample set of gender prediction may include multiple training samples collected during the historical time period. The historical time period can be, for example, the past 7 days, 10 days, and the like. It can be understood that the unknown gender user collected at one time uses the multi-dimensional feature information of the electronic device to form a sample set. The feature set for each sample is recorded with a set of real numbers. The sample values need to be normalized during sampling, such as normalization to a number between 0 and 1. Thus each user is a sample, and N users make up N samples, each of which is characterized by (x ₁ , x ₂ , . . . , x _n ).

The length of time users read a male-like novel;

The length of time users read a female novel;

The length of time users read sports news;

The length of time users read the news of the constellation;

The number of times the user used the front camera to take a self-portrait;

The number of times the user uses the beauty software;

The number and duration of users playing different types of games.

The training module 302 is configured to train the BP neural network model by using the sample set to obtain the trained prediction model.

In an embodiment, the sample parameters may be input into the BP neural network model to obtain the probability of two prediction results, and the probability of obtaining two prediction results based on the first preset formula, wherein the first preset formula is for:

Then, the loss value is obtained according to the two prediction results and the probability corresponding thereto, and the loss value is obtained according to the two prediction results and the probability corresponding thereto according to the second preset formula, wherein the second preset formula is:

Where C is the number of categories of prediction results, y _k is the true value, and E is the average. Then, according to the loss value, training is performed to generate the target model parameters.

The obtaining module 303 is configured to obtain multi-dimensional feature information of the electronic device that the unknown gender user uses and use as a prediction sample.

For example, the user characteristic value mentioned in the previous week is counted, the user feature vector is obtained, and the normalized feature value is input into the network to perform a forward calculation, and the output of the network is the probability that the user gender is male or female. Some of the features described above may be included in the predicted sample.

The generating module 304 is configured to generate a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability comprises: a first probability that the current user is a male, and a second probability that the current user is a female.

In order to simplify the application prediction operation and improve the prediction speed, one probability can be selected for the two probabilities in the prediction probability, and then the gender of the current user is predicted based on the selected probability. For example, for a certain prediction probability, if Y=1 indicates that the current user is male, Y=0 indicates that the current user is female, and if P(Y=1|x) is greater than P(Y=0|x), then the output is The current user is the first predicted result of the male; assuming P(Y=1|x) is not greater than P(Y=0|x), at this time, the second predicted result that the current user is a female is output.

Please refer to FIG. 7. FIG. 7 is another schematic structural diagram of a user gender prediction apparatus according to an embodiment of the present application. In some embodiments, the training module 302 can specifically include a processing sub-module 3021 and a training sub-module 3022.

The processing sub-module 3021 is configured to perform normalization processing on the sample parameters in the sample set;

The training sub-module 3022 is configured to train the BP neural network model according to the normalized sample parameters.

In an embodiment, the training sub-module 3022 is specifically configured to input the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results, according to the two prediction results and corresponding The probability obtains a loss value, and is trained according to the loss value to generate a target model parameter.

In an embodiment, the apparatus 300 may further include:

The comparing module 305 is configured to compare, after the generating module 304 generates the predicted probability according to the predicted sample and the trained predictive model, a first probability that the current user is a male and a second probability that the current user is a female, to obtain a comparison result;

The output module 306 is configured to output a final prediction result according to the comparison result.

Specifically, the output module 306 is specifically configured to: when the first probability is greater than the second probability, output a first prediction result that the current user is a male, when the first probability is not greater than the second probability , output the second predicted result that the current user is a female.

It should be noted that the training process of the predictive model can be completed on the server side or on the electronic device side. When the training process and the actual prediction process of the predictive model are completed on the server side, when the trained predictive model is needed, the multi-dimensional feature information of the electronic device can be input to the server by the user of unknown gender, and after the actual prediction of the server is completed, The prediction result is sent to the electronic device end, and finally the electronic device outputs the predicted result.

It can be seen from the above that the user gender prediction apparatus in the embodiment of the present application constructs a sample set of gender prediction by collecting multi-dimensional feature information of a known gender user using the electronic device, and uses the sample set to train the BP neural network model to obtain The predicted model after the training obtains multi-dimensional feature information of the electronic device of the unknown gender and uses the electronic device as a prediction sample, and generates a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability includes: the first probability that the current user is a male And the second probability that the current user is a woman. The application can predict the gender of the current user based on the BP neural network model, improve the accuracy of the gender prediction, and facilitate the electronic device to complete the accurate portrait of the user.

In the embodiment of the present application, the user gender prediction apparatus belongs to the same concept as the user gender prediction method in the above embodiment, and any method provided in the embodiment of the user gender prediction method may be run on the user gender prediction apparatus, and the specific implementation process thereof For details, refer to the embodiment of the user gender prediction method, and details are not described herein again.

The term "module" as used herein may be taken to mean a software object that is executed on the computing system. The different components, modules, engines, and services described herein can be viewed as implementation objects on the computing system. The apparatus and method described herein may be implemented in software, and may of course be implemented in hardware, all of which are within the scope of the present application.

The embodiment of the present application further provides a storage medium on which a computer program is stored, and when the computer program runs on a computer, causes the computer to execute the user gender prediction method described above.

The embodiment of the present application further provides an electronic device, such as an electronic device such as a tablet computer or a mobile phone. The processor in the electronic device loads the instructions corresponding to the process of one or more applications into the memory according to the following steps, and the processor runs the application stored in the memory to implement various functions:

In an embodiment, when training the BP neural network model with the sample set, the processor is configured to perform the following steps:

Normalizing the sample parameters in the sample set;

In an embodiment, when the BP neural network model is trained according to the normalized sample parameters, the processor is configured to perform the following steps:

In an embodiment, when training is performed according to the loss value, the processor is configured to perform the following steps:

In an embodiment, after generating the prediction probability according to the predicted sample and the trained prediction model, the processor is further configured to perform the following steps:

Based on the comparison result, the final prediction result is output.

In an embodiment, when the final prediction result is output according to the comparison result, the processor is configured to perform the following steps:

References to "an embodiment" herein mean that a particular feature, structure, or characteristic described in connection with the embodiments can be included in at least one embodiment of the present application. The appearances of the phrases in various places in the specification are not necessarily referring to the same embodiments, and are not exclusive or alternative embodiments that are mutually exclusive. Those skilled in the art will understand and implicitly understand that the embodiments described herein can be combined with other embodiments.

Referring to FIG. 8, the electronic device 400 includes a processor 401 and a memory 402. The processor 401 is electrically connected to the memory 402.

The processor 400 is a control center of the electronic device 400 that connects various portions of the entire electronic device using various interfaces and lines, executes the electronic by running or loading a computer program stored in the memory 402, and recalling data stored in the memory 402. The various functions of device 400 and processing data to provide overall monitoring of electronic device 400.

The memory 402 can be used to store software programs and modules, and the processor 401 executes various functional applications and data processing by running computer programs and modules stored in the memory 402. The memory 402 can mainly include a storage program area and a storage data area, wherein the storage program area can store an operating system, a computer program required for at least one function (such as a sound playing function, an image playing function, etc.), and the like; the storage data area can be stored according to Data created by the use of electronic devices, etc. Moreover, memory 402 can include high speed random access memory, and can also include non-volatile memory, such as at least one magnetic disk storage device, flash memory device, or other volatile solid state storage device. Accordingly, memory 402 can also include a memory controller to provide processor 401 access to memory 402.

In the embodiment of the present application, the processor 401 in the electronic device 400 loads the instructions corresponding to the process of one or more computer programs into the memory 402 according to the following steps, and is stored in the memory 402 by the processor 401. The computer program in which to implement various functions, as follows:

The multi-dimensional feature information of the known gender users using the electronic device is collected as a sample, the sample set of gender prediction is constructed, and the BP neural network model is trained by using the sample set to obtain the predicted model after training, and the multi-dimensional use of the electronic device by the unknown gender user is obtained. The feature information is used as a prediction sample, and a prediction probability is generated according to the prediction sample and the trained prediction model, the prediction probability including: a first probability that the current user is a male, and a second probability that the current user is a female. The application can predict the gender of the current user based on the BP neural network model, improve the accuracy of the gender prediction, and facilitate the electronic device to complete the accurate portrait of the user.

Referring to FIG. 9 together, in some embodiments, the electronic device 400 may further include: a display 403, a radio frequency circuit 404, an audio circuit 405, and a power source 406. The display 403, the radio frequency circuit 404, the audio circuit 405, and the power source 406 are electrically connected to the processor 401, respectively.

Display 403 can be used to display information entered by the user or information provided to the user, as well as various graphical user interfaces, which can be comprised of graphics, text, icons, video, and any combination thereof. The display 403 can include a display panel. In some embodiments, the display panel can be configured in the form of a liquid crystal display (LCD), or an organic light-emitting diode (OLED).

The radio frequency circuit 404 can be used to transmit and receive radio frequency signals to establish wireless communication with network devices or other electronic devices through wireless communication, and to transmit and receive signals with network devices or other electronic devices.

The audio circuit 405 can be used to provide an audio interface between the user and the electronic device through the speaker and the microphone.

Power source 406 can be used to power various components of electronic device 400. In some embodiments, the power supply 406 can be logically coupled to the processor 401 through a power management system to enable functions such as managing charging, discharging, and power management through the power management system.

Although not shown in FIG. 9, the electronic device 400 may further include a camera, a Bluetooth module, and the like, and details are not described herein.

In the embodiment of the present application, the storage medium may be a magnetic disk, an optical disk, a read only memory (ROM), or a random access memory (RAM).

In the above embodiments, the descriptions of the various embodiments are different, and the details that are not detailed in a certain embodiment can be referred to the related descriptions of other embodiments.

It should be noted that, for the user gender prediction method in the embodiment of the present application, a general tester in the field can understand all or part of the process of implementing the user gender prediction method in the embodiment of the present application, and the related hardware can be controlled by a computer program. Upon completion, the computer program can be stored in a computer readable storage medium, such as in a memory of the electronic device, and executed by at least one processor within the electronic device, and can include, for example, a user gender prediction method during execution. The flow of the embodiment. The storage medium may be a magnetic disk, an optical disk, a read only memory, a random access memory, or the like.

For the user gender prediction apparatus of the embodiment of the present application, each functional module may be integrated into one processing chip, or each module may exist physically separately, or two or more modules may be integrated into one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules. An integrated module, if implemented in the form of a software functional module and sold or used as a standalone product, may also be stored in a computer readable storage medium such as a read only memory, a magnetic disk or an optical disk.

The user gender prediction method, apparatus, storage medium, and electronic device provided by the embodiments of the present application are described in detail. The principles and implementation manners of the present application are described in the specific examples, and the foregoing embodiments are described. It is only used to help understand the method of the present application and its core idea; at the same time, for those skilled in the art, according to the idea of the present application, there will be changes in the specific implementation manner and application scope. The contents of the description should not be construed as limiting the application.

Claims

A user gender prediction method, wherein the method comprises the following steps:

Collecting multi-dimensional feature information of a known gender user using the electronic device as a sample, and constructing a sample set of gender prediction;

Using the sample set to train the BP neural network model to obtain a trained prediction model;

Obtaining multi-dimensional feature information of an electronic device that is used by an unknown gender user as a prediction sample;

A prediction probability is generated according to the predicted sample and the trained prediction model, the predicted probability including: a first probability that the current user is a male, and a second probability that the current user is a female.
The user gender prediction method according to claim 1, wherein the step of training the BP neural network model by using the sample set comprises:

Normalizing the sample parameters in the sample set;

The BP neural network model is trained according to the normalized sample parameters.
The user gender prediction method according to claim 2, wherein the step of training the BP neural network model according to the normalized sample parameters comprises:

Entering the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results;

Obtaining a loss value according to the two prediction results and a probability corresponding thereto;

Training is performed based on the loss value to generate a target model parameter.
The user gender prediction method according to claim 3, wherein the step of training according to the loss value comprises:

Training is performed using the gradient descent method based on the loss value.
The user gender prediction method according to claim 1, wherein after generating the prediction probability according to the prediction sample and the trained prediction model, the method further comprises:

Comparing a first probability that the current user is a male with a second probability that the current user is a female, and obtaining a comparison result;

Based on the comparison result, the final prediction result is output.
The user gender prediction method according to claim 5, wherein the step of outputting the final prediction result according to the comparison result comprises:

When the first probability is greater than the second probability, outputting a first prediction result that the current user is a male;

When the first probability is not greater than the second probability, outputting a second prediction result that the current user is a female.
The user gender prediction method according to claim 3, wherein the step of inputting the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results comprises:

Calculating a probability based on a first preset formula to obtain two prediction results, wherein the first preset formula is:

Where Z K is the intermediate value and C is the number of categories of the predicted result.
For the jth intermediate value, a two-dimensional output layer is obtained based on the first preset formula, respectively representing the probability that the sample is male and female.
The user gender prediction method according to claim 3, wherein the step of obtaining a loss value according to the two prediction results and the probability corresponding thereto comprises:

And obtaining, according to the second preset formula, a loss value according to the two prediction results and a probability corresponding thereto, wherein the second preset formula is:

Where C is the number of categories of prediction results, y k is the true value, and E is the average.
A user gender prediction device, wherein the device comprises:

An acquisition module is configured to collect multi-dimensional feature information of a known gender user using the electronic device as a sample, and construct a sample set of gender prediction;

a training module, configured to train the BP neural network model by using the sample set to obtain a trained prediction model;

An obtaining module, configured to obtain multi-dimensional feature information of an electronic device used by an unknown gender user as a prediction sample;

And a generating module, configured to generate a prediction probability according to the predicted sample and the trained prediction model, where the predicted probability includes: a first probability that the current user is a male, and a second probability that the current user is a female.
The user gender prediction apparatus according to claim 9, wherein the training module specifically includes:

Processing a sub-module for normalizing the sample parameters in the sample set;

The training submodule is configured to train the BP neural network model according to the normalized sample parameters.
The user gender prediction apparatus according to claim 10, wherein

The training sub-module is specifically configured to input the normalized sample parameters into the BP neural network model to obtain a probability of two prediction results, and obtain a loss value according to the two prediction results and a probability corresponding thereto, according to The loss value is trained to generate a target model parameter.
The user gender prediction apparatus according to claim 9, wherein the apparatus further comprises:

a comparison module, configured to compare a first probability that the current user is a male with a second probability that the current user is a female, and obtain a comparison result, after the generating module generates the predicted probability according to the predicted sample and the trained predicted model ;

And an output module, configured to output a final prediction result according to the comparison result.
The user gender prediction apparatus according to claim 12, wherein

The output module is configured to: when the first probability is greater than the second probability, output a first prediction result that the current user is a male, and when the first probability is not greater than the second probability, output the current The user is the second predicted result of the woman.
A storage medium having stored thereon a computer program, wherein when the computer program is run on a computer, the computer is caused to perform the user gender prediction method according to any one of claims 1 to 8.
An electronic device includes a processor and a memory, the memory storing a plurality of instructions, wherein the processor loads instructions in the memory for performing the following steps:

Collecting multi-dimensional feature information of a known gender user using the electronic device as a sample, and constructing a sample set of gender prediction;

Using the sample set to train the BP neural network model to obtain a trained prediction model;

Obtaining multi-dimensional feature information of an electronic device that is used by an unknown gender user as a prediction sample;

A prediction probability is generated according to the predicted sample and the trained prediction model, the predicted probability including: a first probability that the current user is a male, and a second probability that the current user is a female.
The electronic device of claim 15, wherein when the BP neural network model is trained using the sample set, the processor is configured to perform the following steps:

Normalizing the sample parameters in the sample set;

The BP neural network model is trained according to the normalized sample parameters.
The electronic device of claim 16, wherein the processor is configured to perform the following steps when training the BP neural network model based on the normalized sample parameters:

Entering the normalized sample parameters into the BP neural network model to obtain the probability of two prediction results;

Obtaining a loss value according to the two prediction results and a probability corresponding thereto;

Training is performed based on the loss value to generate a target model parameter.
The electronic device of claim 17, wherein said processor is operative to perform the following steps when training is performed based on said loss value:

Training is performed using the gradient descent method based on the loss value.
The electronic device of claim 15, wherein the processor is further configured to perform the following steps after generating the prediction probability according to the predicted sample and the trained prediction model:

Comparing a first probability that the current user is a male with a second probability that the current user is a female, and obtaining a comparison result;

Based on the comparison result, the final prediction result is output.
The electronic device according to claim 19, wherein, when the final prediction result is output according to the comparison result, the processor is configured to perform the following steps:

When the first probability is greater than the second probability, outputting a first prediction result that the current user is a male;

When the first probability is not greater than the second probability, outputting a second prediction result that the current user is a female.