WO2021175010A1 - Appareil et procédé d'identification de sexe, dispositif électronique et support de stockage - Google Patents

Appareil et procédé d'identification de sexe, dispositif électronique et support de stockage Download PDF

Info

Publication number
WO2021175010A1
WO2021175010A1 PCT/CN2021/070935 CN2021070935W WO2021175010A1 WO 2021175010 A1 WO2021175010 A1 WO 2021175010A1 CN 2021070935 W CN2021070935 W CN 2021070935W WO 2021175010 A1 WO2021175010 A1 WO 2021175010A1
Authority
WO
WIPO (PCT)
Prior art keywords
gender
data
user
interaction
result
Prior art date
Application number
PCT/CN2021/070935
Other languages
English (en)
Chinese (zh)
Inventor
牛姣姣
Original Assignee
深圳壹账通智能科技有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 深圳壹账通智能科技有限公司 filed Critical 深圳壹账通智能科技有限公司
Publication of WO2021175010A1 publication Critical patent/WO2021175010A1/fr

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/958—Organisation or management of web site content, e.g. publishing, maintaining pages or automatic linking
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90—Details of database functions independent of the retrieved data types
    • G06F16/95—Retrieval from the web
    • G06F16/953—Querying, e.g. by the use of web search engines
    • G06F16/9535—Search customisation based on user profiles and personalisation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions
    • G06Q30/0601—Electronic shopping [e-shopping]
    • G06Q30/0631—Item recommendations

Definitions

  • This application relates to the field of big data technology, and in particular to a method, device, electronic device, and computer-readable storage medium for user gender identification.
  • service providers use terminal devices to provide services to users.
  • service providers In the process of users receiving services, service providers often collect information on the user's gender, age, occupation and other aspects to provide a basis for personalized services.
  • the terminal device obtains the user's face image by taking an image or receiving an image uploaded by the user, and then extracts feature data from the face image to identify the gender of the user based on the feature data of the face image.
  • the user's facial image collection requires the user to make actions such as shaking his head, blinking to cooperate with the shooting, or requires the user to upload a face image that meets specific requirements, and interacts with the user more and more. It is complicated, and it is easy for users to repeat operations due to poor natural conditions such as venues and light, or poor imaging equipment capabilities, and require users to repeat operations, resulting in cumbersome operations and low operating efficiency in the user's gender identification process.
  • This application provides a user gender identification method, including: acquiring interaction data in a user interaction process performed by a client, the interaction data including interaction behavior data and interaction time data; by calculating the interaction behavior data and pre-collected data The correlation between the sets is determined to determine the first gender result reflected by the interactive behavior data, the data set contains historical interactive behavior data collected for users who determine the gender, and gender analysis is performed on the interaction time data , Obtain the second gender result reflected by the interaction time data; determine the gender of the user interacting with the user in the client according to the first gender result and the second gender result.
  • the present application also provides a user gender identification device, including: an interaction data acquisition module for acquiring interaction data in a user interaction process performed by the client, the interaction data including interaction behavior data and interaction time data; gender result determination module , Used to determine the first gender result reflected by the interactive behavior data by calculating the correlation between the interactive behavior data and the pre-collected data set, the data set containing the history collected for the user of the determined gender Interactive behavior data, and used to obtain the second gender result reflected by the interactive time data by performing gender analysis on the interactive time data; the user gender determination module is used to obtain the second gender result reflected by the interaction time data according to the first gender result and the first gender result The second gender result is to determine the gender of the user interacting with the client in the client.
  • an interaction data acquisition module for acquiring interaction data in a user interaction process performed by the client, the interaction data including interaction behavior data and interaction time data
  • gender result determination module Used to determine the first gender result reflected by the interactive behavior data by calculating the correlation between the interactive behavior data and the pre-collected
  • the present application also provides an electronic device, including a processor and a memory, and computer-readable instructions are stored on the memory.
  • a processor When the computer-readable instructions are executed by the processor, the following steps are implemented:
  • interaction data in the user interaction process performed by the client including interaction behavior data and interaction time data; determine the interaction behavior by calculating the correlation between the interaction behavior data and the pre-collected data set
  • the first gender result reflected by the data the data set contains historical interaction behavior data collected for users with a certain gender
  • the second gender reflected by the interaction time data is obtained by performing gender analysis on the interaction time data Gender result; according to the first gender result and the second gender result, the gender of the user interacting with the user in the client is determined.
  • This application also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, the following steps are implemented:
  • interaction data in the user interaction process performed by the client including interaction behavior data and interaction time data; determine the interaction behavior by calculating the correlation between the interaction behavior data and the pre-collected data set
  • the first gender result reflected by the data the data set contains historical interaction behavior data collected for users with a certain gender
  • the second gender reflected by the interaction time data is obtained by performing gender analysis on the interaction time data Gender result; according to the first gender result and the second gender result, the gender of the user interacting with the user in the client is determined.
  • Figure 1 is a schematic diagram of an implementation environment involved in this application.
  • Fig. 2 is a flow chart showing a method for user gender identification according to an exemplary embodiment
  • FIG. 3 is a flowchart of one embodiment of step 203 in the embodiment corresponding to FIG. 2;
  • FIG. 4 is a flowchart of one embodiment of step 301 in the embodiment corresponding to FIG. 3;
  • FIG. 5 is a flowchart of one embodiment of step 203 in the embodiment corresponding to FIG. 2;
  • FIG. 6 is a flowchart of one embodiment of step 205 in the embodiment corresponding to FIG. 2;
  • FIG. 7 is a flowchart of another embodiment of step 205 in the embodiment corresponding to FIG. 2;
  • Fig. 8 is a block diagram showing a device for identifying user gender according to an exemplary embodiment
  • Fig. 9 shows a schematic diagram of the hardware structure of an electronic device according to an exemplary embodiment.
  • Fig. 1 is a schematic diagram of an implementation environment involved in this application.
  • the implementation environment may include: a terminal 101 and a server 103.
  • the terminal 101 establishes a wired or wireless communication connection with the server 103 in advance, and then realizes data transmission with the server 103 through this communication connection.
  • a client terminal is running in the terminal 101, and the client terminal is used to provide a user interaction interface to provide data services for the user according to the user interaction operation triggered in the user interaction interface.
  • the client may be a shopping client.
  • personalized product recommendations can be made for the user according to the user's gender.
  • the terminal 101 may be a mobile phone, a tablet computer, a notebook computer, a smart teller machine, or any other electronic device that can be operated by the client, and there is no restriction here.
  • the client can be an application client or a web client, and there are no restrictions here.
  • the server 103 is used to provide data support for user interaction operations triggered in the client, so that the client can run normally in the terminal 101. Still taking the above shopping client as an example, the server 103 can provide the shopping client with product detail data according to the product detail browsing operation triggered in the shopping client, and can also identify the user’s gender so that when the user triggers the product browsing operation Provide the shopping client with products that match the user’s gender.
  • the server 103 may be a single server device or a server group composed of multiple server devices, which is not limited here.
  • Fig. 2 is a flow chart showing a method for user gender identification according to an exemplary embodiment.
  • the method may be executed by the server 103 in the implementation environment shown in Fig. 1.
  • the method may include the following steps:
  • Step 201 Obtain interaction data during a user interaction process performed by the client, where the interaction data includes interaction behavior data and interaction time data.
  • the user is required to make actions such as shaking his head and blinking to cooperate with the face collection, and it is easy to cause recognition failure due to poor natural conditions such as the venue and light, or poor imaging equipment capabilities, and the user is required to repeat
  • the operation results in a very complicated process of gender recognition and the recognition efficiency is not high.
  • this embodiment proposes a user gender recognition method, which can effectively recognize the user’s gender without collecting the user’s face image, thereby effectively avoiding the dependence on the user’s face recognition in the prior art. The problem caused.
  • the user interaction process performed by the client refers to the process in which the user performs user operations based on the user interaction interface provided by the client.
  • the user operations performed by the user are user interaction behaviors, which are exemplary,
  • the user's interaction behavior may include operations such as single-click, double-click, long-press, and slide.
  • the interaction data in the user interaction process is data related to user operations performed on the user interaction interface, for example, may include interaction behavior data related to the user's interaction behavior, and may also include interaction time data related to the user's interaction time.
  • the interactive behavior data may include data such as the service type of the data service enabled by the user in the client, the user's keystroke area, keystroke pressure, keystroke position or trajectory, and keystroke time.
  • the interaction time data may include data related to the duration of the user's interaction behavior, such as the duration of the user's continuous use of a data service of a certain service type, and the total duration of the user's use of the client in a period of time.
  • Step 203 Determine the first gender result reflected by the interaction behavior data by calculating the correlation between the interaction behavior data and the pre-collected data set.
  • the data set contains historical interaction behavior data collected for the user of the determined gender, and through Gender analysis is performed on the interaction time data, and the second gender result reflected by the interaction time data is obtained.
  • the correlation between the interaction behavior data obtained in step 201 and the historical interaction behavior data in the data set can be obtained by calculating the correlation between the interaction behavior data obtained in step 201 and the historical interaction behavior data in the data set.
  • the first gender result can reflect the gender of the user currently performing user interaction in the client from the aspect of the user's interaction behavior.
  • the result of the first gender obtained reflects the possibility that the gender of the user currently performing user interaction in the client is male; if the data collected in the data set The historical interaction behavior data all correspond to female users, and the first gender result reflects the possibility that the gender of the user currently performing user interaction in the client is female.
  • the first gender result reflects the possibility that the genders of the users currently performing user interactions in the client are male and female respectively.
  • Interaction time data reflects the duration of user interaction behavior, which can reflect user gender to a certain extent. For example, female users usually spend more time on shopping clients than male users, and male users usually spend more time browsing current affairs news clients than female users, so you can pass Perform gender analysis on the interaction time data obtained in step 201, and obtain the possibility that the gender of the user currently performing user interaction in the client is male or female.
  • the second gender result reflects the gender of the user currently performing user interaction in the client in terms of the user's interaction time.
  • Step 205 Determine the gender of the user interacting with the user in the client according to the first gender result and the second gender result.
  • the first gender result can reflect the gender of the user currently performing user interaction in the client in terms of interaction behavior
  • the second gender result can reflect the gender of the user currently performing user interaction in the client in terms of interaction behavior. Therefore, this implementation For example, by combining user interaction behavior and user interaction time to determine the gender of the user interacting with the client, the user gender can be accurately obtained.
  • the probability of belonging to the same gender in the first gender result and the second gender result may be calculated as an average value or a weighted sum, and the gender corresponding to a larger probability is determined as the user interacting in the client.
  • User gender may be calculated as an average value or a weighted sum, and the gender corresponding to a larger probability is determined as the user interacting in the client.
  • the second gender result indicates that the probability of the gender of the user interacting in the client is male is 80%, and the probability of being female is 20 %, it can be determined that the gender of the user interacting with the client is male.
  • the first gender result obtained by calculating the correlation between the interactive behavior data and the data set reflects the current user gender Correlation with the user's determined gender in the data set, so the first gender result reflects the current user's gender in terms of interactive behavior.
  • the gender of the current user can be judged based on the characteristics of the interaction time. Therefore, the second gender result reflects the gender of the current user in terms of the interaction time. Therefore, this application combines two aspects of user interaction behavior and user interaction time to jointly determine the user's gender.
  • the user’s gender is no longer determined based on face recognition, but the user’s gender is determined based on the user’s interactive behavior data in the client, without the user’s active cooperation in shooting or uploading images, and at the same time It avoids repeated shooting or uploading caused by failure of recognition due to problems such as venue, light or imaging equipment, and completes gender recognition without the user feeling, which greatly improves the efficiency of user gender recognition.
  • determining the first gender result reflected by the interaction behavior data may include the following steps:
  • Step 301 Normalize the interactive behavior data according to a preset gender data standard to obtain first standard data.
  • the gender data standard is used to describe the processing standard for normalizing the interactive behavior data.
  • the gender data standard is the standard reference data for normalizing the interactive behavior data set for gender analysis.
  • the gender data standard may include a data attribute table related to gender analysis, for example, a male score and a female score corresponding to a data service of a specific business type are recorded, and the average keystroke area and average keystroke pressure corresponding to male and female respectively , Common keystroke position trajectory and standard parameter data such as average keystroke time.
  • gender data standards can be set for different genders in advance to obtain user interaction behavior data. Then, select the appropriate gender data standard to process this interactive behavior data.
  • the normalization process is a process of standardizing the interaction behavior data according to a preset gender data standard.
  • the normalization process may be a process of converting the interaction behavior data into a corresponding feature vector.
  • the obtained first standard data may be corresponding to a preset gender and related to the user interaction behavior. Eigenvectors.
  • data that is useless for gender analysis in the interaction behavior data can also be removed in advance, or redundant data in the interaction behavior data, such as data that repeatedly appears multiple times. .
  • redundant data in the interaction behavior data such as data that repeatedly appears multiple times.
  • Step 303 Calculate the Spearman correlation coefficient and the Pearson correlation coefficient between the first standard data and the second standard data corresponding to the user whose gender is determined, respectively.
  • the second standard data is based on the gender data standard corresponding to the determined gender.
  • the historical interactive behavior data is normalized.
  • the second standard data is obtained by normalizing the historical interaction behavior data of users whose gender is determined in the data set according to the gender data standard corresponding to the determined gender in the data set. Therefore, the second standard data is the same as the first standard data.
  • the data are all related to gender data standards.
  • Spearman's correlation coefficient and Pearson's correlation coefficient are used to measure the degree of correlation between the first standard data and the second standard data, because the user gender corresponding to the second standard data is known Therefore, according to the degree of correlation between the first standard data and the second standard data, the degree of correlation between the gender of the user interacting with the user in the client and the known user gender can be obtained, and the current The gender of the user who interacted with the user.
  • N represents the type of data contained in the interactive behavior data, for example, the interactive behavior data only contains the keystroke area and keystroke pressure
  • the value of N is 2
  • X represents the first standard data
  • Y Represents the second standard data.
  • the Pearson correlation coefficient can be divided into three levels.
  • ⁇ 0.4 it means that the gender of the user interacting with the user in the client is more different from the determined user gender.
  • ⁇ 0.7 means that the gender of the user interacting in the client is more different from the determined user gender.
  • the determined user gender is highly linearly related.
  • the i-th (1 ⁇ i ⁇ M) data randomly selected from these two sets of standard data are used as X i , Y i means, and then sort X and Y in ascending or descending order to get the sorted X'and Y'.
  • the data set contains historical interactive behavior data corresponding to males and historical interactive behavior data corresponding to females, respectively, it is necessary to calculate the Spearman correlation coefficients and the correlation coefficients between the first standard data and the second standard data corresponding to users of different genders. Pearson's correlation coefficient.
  • Step 305 Obtain the first gender result by calculating the weighted sum of Spearman's correlation coefficient and Pearson's correlation coefficient.
  • the weight parameter corresponding to Spearman's correlation coefficient r is 0.6
  • the weight parameter corresponding to Pearson's correlation coefficient ⁇ is 0.4
  • the weighted sum of Spearman's correlation coefficient and Pearson's correlation coefficient is 0.6*r+0.4* ⁇ .
  • weight parameters are for example purposes only, and the actual weight values can be adjusted based on experience or actual conditions.
  • the first gender result obtained by calculating the weighted sum of Spearman's correlation coefficient and Pearson's correlation coefficient indicates that the gender of the user interacting in the client is this Probability of single sex.
  • the weighted sum of Spearman's correlation coefficient and Pearson's correlation coefficient must be calculated for different genders to obtain the first gender result It includes the probability that the gender of the user interacting with the user in the client is of a different gender.
  • the interaction behavior data is normalized and the Spearman correlation coefficient and the Pearson correlation coefficient are two different correlation coefficients used to calculate the interaction behavior data and history.
  • the relevance of interactive behavior data can effectively ensure the accuracy of the first gender results.
  • the interaction behavior data is normalized according to a preset gender data standard, and the process of obtaining the first standard data may include the following steps:
  • Step 401 Determine a gender data standard that matches the service type according to the service type triggered in the client to enable the data service;
  • Step 403 Perform normalization processing on the interaction behavior data according to the processing standard described in the determined gender data standard to obtain first standard data.
  • gender data standards can include male scores and female scores corresponding to data services of a specific business type. Therefore, the gender that matches this business type can be determined according to the business type of the data service triggered in the client. Data standards.
  • the gender data standard can be determined as the gender data standard corresponding to the female. If the current affairs news browsing is triggered in the client, the gender data standard can be determined as the gender data standard corresponding to males.
  • the corresponding gender data standard will be obtained according to the service type of the data service activated in the interactive behavior data, and the interactive behavior data will be normalized according to the gender data standard to obtain accurate The first standard data to lay the foundation for the accuracy of subsequent user gender identification.
  • the process of obtaining the second gender result reflected by the interaction time data may include the following steps:
  • Step 501 Extract time-domain features in the user interaction process according to the interaction time data, where the time-domain features include at least one of periodicity, rate of change, and acceleration.
  • the interaction time data reflects the duration of the user interaction behavior, so the interaction time data should contain the time information in the user interaction process. By extracting the characteristics of these time information, the time domain in the user interaction process can be obtained. feature.
  • the time-domain features extracted from the interaction time data may include at least one of periodicity, rate of change, and acceleration.
  • the extracted periodicity may include the periodicity of the user browsing a certain product (such as women’s clothing, etc.), the rate of change in the length of browsing time, and the process of switching shopping pages. The acceleration of the user's sliding operation, etc.
  • Step 503 Input the temporal characteristics into a preset gender determination model, so that the gender determination model predicts the gender of the user interacting with the user in the client according to the temporal characteristics, and obtains a second gender result.
  • the gender determination model is a machine learning model used to predict the user's gender in the client for user interaction based on the temporal characteristics in the input model.
  • it can be a supervised machine learning model or a semi-supervised machine learning model.
  • a supervised machine learning model or a semi-supervised machine learning model.
  • a semi-supervised machine learning model Including support vector machine (Support Vector Machine, SVM) model, support vector clustering (support vector clustering) model or semi-supervised SVM (Semi-Supervised SVM, S3VM) model, etc.
  • the gender determination model uses the time-domain feature data that has been marked with gender and combines it with the unlabeled time-domain feature data for training, so that the model converges and the weight parameters in the model are optimized, so that the model can be applied in practice.
  • Step 505 Obtain the second gender result output by the gender determination model.
  • the second gender result may include the probability that the gender of the user currently interacting with the user in the client is male or female, and may also include the probability of male and female respectively.
  • the time domain features in the user interaction process are extracted according to the interaction time data, and the pre-trained gender determination model is used to identify the time domain features.
  • the gender identification can be performed from the user's interaction time. Determining the gender of the user provides a further basis to further improve the accuracy of gender identification.
  • the historical interaction behavior data contained in the data set corresponds to a single determined gender, that is, the historical interaction behavior data contained in the data set only corresponds to male users or female users. Based on the foregoing, it can be known that in this case
  • the first gender result describes the probability that the gender of the user interacting with the user in the client corresponds to a single determined gender.
  • determining the gender of the user interacting with the user in the client may include the following steps:
  • Step 601 Calculate the weighted sum of the first gender result and the second gender result.
  • Step 603 If the value corresponding to the weighted sum is above the first threshold, it is determined that the user's gender is the same as the single determined gender, where the first threshold is set in advance for the user of the single determined gender.
  • the first threshold is set in advance for a user with a single determined gender, and is used to determine the minimum credibility that the identification result of the user's gender is the same as the single determined gender. For example, the first threshold is set for an authorized user of the client, and the authorized user collects the user's gender information during authorization, so the user's gender corresponding to the first threshold is also determined.
  • the weighted sum of the first gender result and the second gender result is greater than the first threshold, it means that the gender of the user currently interacting with the user in the client is the same as the single determined gender. For example, if the weighted sum is calculated to indicate that the probability that the user's gender is male is 70%, and the first threshold is 60%, then the user's gender can be male.
  • the client can send service recommendation information to the client according to the determined user gender, and the client can then recommend the user to the user based on the service recommendation information.
  • Gender-compatible business after determining the gender of the user interacting with the user in the client, the client can send service recommendation information to the client according to the determined user gender, and the client can then recommend the user to the user based on the service recommendation information. Gender-compatible business.
  • the client may be an unauthorized user
  • the client can provide risk warnings to authorized users based on this information, for example, send risk warning messages or emails to authorized users according to the contact information reserved by the authorized users.
  • an alarm threshold may also be set, the alarm threshold is less than or equal to the first threshold, and in the case that the weighted sum is less than the first threshold, the weighted sum is compared with the alarm threshold. If the weighted sum is less than the alarm threshold, an SMS prompt is sent to the authorized user, or the user sends a verification code and asks the user to enter the verification code on the client to continue the operation. If the weighted sum is greater than the alarm threshold, only risk warning is performed on the page, so that the alarm threshold can be used to perform different levels of risk control for different gender identification results, which can reduce the number of false alarms and improve system efficiency.
  • the first gender result is calculated for a single determined gender, and the weighted sum of the first gender result and the second gender result is compared with the threshold for the single determined gender to determine whether the user belongs to that gender.
  • the single determination of gender can simplify the calculation complexity of the gender identification process, save computing resources, reduce implementation costs, and improve the security of business services.
  • the data set contains historical interaction behavior data collected for users of different genders
  • the first gender result includes the correlation between the interaction behavior data and the historical interaction behavior data of different users. The result of determining the sex of the child.
  • the data set contains historical interaction behavior data corresponding to multiple users
  • the first gender result includes the correlation between the gender of the user interacting with the user in the client and the gender of each user in the data set.
  • determining the gender of the user interacting with the user in the client according to the first gender result and the second gender result may include the following steps:
  • Step 701 Calculate the weighted sum of each sub-gender result and the second gender result respectively, and determine the user gender whose weighted sum is greater than a preset second threshold according to the data set.
  • the calculation method of the weighted sum of each sub-gender result and the second gender result is the same as the calculation of the weighted sum of the first gender result and the second gender result in the embodiment shown in FIG. 6.
  • the same second threshold is used for judgment for all weighted sums of each sub-gender result and the second gender result. For example, if the data set contains historical interaction behavior data collected for two male users A and B and two female users C and D, the first gender result includes four sub-gender results, each of which is associated with the second gender result Four weighted sums are calculated. Among them, the weighted sum calculated with A is 70%, the weighted sum calculated with B is 80%, the weighted sum calculated with C is 30%, and the weighted sum calculated with D is 60%. If the same second threshold value is set to 50% for all weighted sums, it is determined that the gender of the user whose weighted sum is greater than the second threshold is two males and one female.
  • different second thresholds may be used for each weighted sum of each sub-gender result and the second gender result, or different second thresholds may be used for different genders, which is not limited here.
  • Step 703 Determine a user gender with a larger number of weighted sums greater than a preset second threshold as the user gender who performs user interaction in the client.
  • the weighted sum number of user genders greater than the preset second threshold is larger, it means that the user gender interacting with the user in the client is more likely to be the user gender with more data, so that the user gender is Determined as the gender of the user currently interacting with the client.
  • the gender of the user whose weighted sum is greater than the second threshold is two males and one female, and at this time, it is determined that the gender of the user interacting with the user in the client is male.
  • the number of male users and female users that are greater than the weighted sum of the preset second threshold may be the same.
  • the weighted sums greater than the preset second threshold may be summed separately for different genders, and the sex with the larger sum result is determined as the user gender. For example, if two weighted sums corresponding to males are calculated to be 70% and 80%, and two weighted sums corresponding to females are 60% and 60%, then these four weighted sums can be summed for different genders.
  • the sum result corresponding to the male is 150%.
  • the sum corresponding to the female is 120%, and thus the gender of the user interacting with the client in the client is determined to be male.
  • multiple sub-gender results are obtained by calculating the correlation coefficient between the current user's interactive behavior data and multiple users' historical interactive behavior data, and the multiple sub-gender results and the second gender result are combined to determine User gender can reduce the difference between current user interaction behavior data and historical interaction behavior data caused by individual differences of users, and improve the accuracy of gender recognition results.
  • Fig. 8 is a block diagram showing a user gender recognition device according to an exemplary embodiment.
  • the user gender identification device includes an interactive data acquisition module 810, a gender result determination module 830, and a user gender determination module 850.
  • the interaction data acquisition module 810 is configured to acquire interaction data in a user interaction process performed by the client, and the interaction data includes interaction behavior data and interaction time data.
  • the gender result determination module 830 is used to determine the first gender result reflected by the interaction behavior data by calculating the correlation between the interaction behavior data and the pre-collected data set, where the data set contains historical interactions collected for the user of the determined gender Behavioral data, and through gender analysis of the interaction time data, the second gender result reflected by the interaction time data is obtained.
  • the user gender determination module 850 is configured to determine the gender of the user interacting with the user in the client according to the first gender result and the second gender result.
  • the gender result determination module 830 includes a normalization unit, a correlation coefficient calculation unit, and a first gender result acquisition unit.
  • the normalization unit is used to normalize the interactive behavior data according to a preset gender data standard to obtain first standard data, where the gender data standard is used to describe the processing standard for normalizing the interactive behavior data.
  • the correlation coefficient calculation unit is used to calculate the Spearman correlation coefficient and the Pearson correlation coefficient between the first standard data and the second standard data corresponding to the user whose gender is determined, wherein the second standard data corresponds to the determined gender Gender data standard, obtained by normalizing historical interactive behavior data.
  • the first gender result obtaining unit is used to obtain the first gender result by calculating the weighted sum of Spearman's correlation coefficient and Pearson's correlation coefficient.
  • the normalization unit includes a data standard determination subunit and a normalization subunit.
  • the data standard determining subunit is used to determine the gender data standard matching the service type according to the service type triggered in the client to enable the data service.
  • the normalization subunit is used to normalize the interactive behavior data according to the processing standard described in the determined gender data standard to obtain the first standard data.
  • the gender determination module 830 includes a time domain feature extraction unit, a time domain feature input unit, and a second gender result acquisition unit.
  • the time-domain feature extraction unit is configured to extract time-domain features in the user interaction process according to the interaction time data, where the time-domain features include at least one of periodicity, rate of change, and acceleration.
  • the time-domain feature input unit is used to input the time-domain feature into a preset gender determination model, so that the gender-determination model predicts the gender of the user interacting with the client in the client according to the time-domain feature, and obtains a second gender result.
  • the second gender result obtaining unit is used to obtain the second gender result output by the gender determination model.
  • the historical interaction behavior data contained in the data set corresponds to a single determined gender
  • the first gender result describes the probability that the gender of the user interacting with the client in the client corresponds to the single determined gender; the user gender is determined
  • the module 850 includes a first weighted sum calculation unit and a first gender judgment unit.
  • the first weighted sum calculation unit is used to calculate the weighted sum of the first gender result and the second gender result.
  • the first gender judging unit is configured to determine that the user's gender is the same as the single determined gender when the weighted sum corresponding value is above the first threshold, wherein the first threshold is set in advance for the user of the single determined gender.
  • the first gender judging unit is further configured to control the client to take risk response when the value corresponding to the weighted sum is below the first threshold.
  • the data set contains historical interaction behavior data collected for users of different genders
  • the first gender result includes the correlation between the interaction behavior data and the historical interaction behavior data of different users, respectively.
  • the determined sub-gender result; the user gender determination module 850 includes a second weighted sum calculation unit and a second gender judgment unit.
  • the second weighted sum calculation unit is configured to calculate the weighted sum of each sub-gender result and the second gender result, and determine the user gender whose weighted sum is greater than the preset second threshold according to the data set.
  • the second gender judging unit is configured to determine the gender of the user whose weighted sum is larger than the preset second threshold value as the gender of the user interacting with the user in the client terminal.
  • the present application further provides an electronic device, including: a processor; and a memory, wherein a computer readable instruction is stored in the memory, and the computer readable instruction is executed by the processor to realize each of the foregoing.
  • the user gender identification method in the embodiment is not limited to: a processor; and a memory, wherein a computer readable instruction is stored in the memory, and the computer readable instruction is executed by the processor to realize each of the foregoing.
  • Fig. 9 is a schematic diagram showing a hardware structure of an electronic device according to an exemplary embodiment.
  • this device is only an example adapted to this application, and cannot be considered as providing any limitation on the scope of use of this application.
  • the device also cannot be interpreted as needing to rely on or have one or more components in the exemplary electronic device shown in FIG. 9.
  • the device includes: a power supply 1100, an interface 1300, at least one memory 1500, and at least one central processing unit (CPU, Central Processing Units) 1700.
  • the power supply 1100 is used to provide working voltage for each hardware device on the device.
  • the interface 1300 includes at least one wired or wireless network interface 1310, at least one serial-to-parallel conversion interface 1330, at least one input/output interface 1350, at least one USB interface 1370, etc., for communicating with external devices.
  • the memory 1500 may be volatile or non-volatile.
  • the memory 1500 as a resource storage carrier, can be a read-only memory, a random access memory, a magnetic disk or an optical disc, etc.
  • the resources stored on it include the operating system 1510, application programs 1530 or data 1550, etc.
  • the storage method can be short-term storage or permanent storage.
  • the operating system 1510 is used to manage and control various hardware devices and application programs 1530 on the device to realize the calculation and processing of massive data 1550 by the central processing unit 1570, which can be Windows ServerTM, Mac OS XTM, UnixTM, LinuxTM Wait.
  • the application program 1530 is a computer program that completes at least one specific task based on the operating system 1510. It may include at least one module, and each module may respectively contain a series of computer-readable instructions for the device.
  • the central processing unit 1570 may include one or more processors, and is configured to communicate with the memory 1500 through a bus, and is used for computing and processing the massive data 1550 in the memory 1500.
  • the electronic device applicable to the present application will complete the aforementioned user gender identification method through the central processing unit 1700 reading a series of computer-readable instructions stored in the memory 1500.
  • this application can also be implemented by hardware circuits or hardware circuits in combination with software instructions. Therefore, implementation of this application is not limited to any specific hardware circuits, software, and combinations of the two.
  • a computer-readable storage medium is also provided.
  • the computer-readable storage medium may be volatile or non-volatile.
  • a computer program is stored thereon, and when the computer program is executed by a processor, the user gender identification method in the foregoing embodiments is realized.

Landscapes

  • Engineering & Computer Science (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • General Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Economics (AREA)
  • General Business, Economics & Management (AREA)
  • Strategic Management (AREA)
  • Marketing (AREA)
  • Development Economics (AREA)
  • User Interface Of Digital Computer (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

La présente invention concerne un procédé et un appareil d'identification de sexe, un dispositif électronique et un support de stockage lisible par ordinateur. Le procédé consiste à : acquérir des données d'interaction pendant un processus d'interaction d'utilisateur qui est réalisé chez un client, les données d'interaction comprenant des données de comportement d'interaction et des données de temps d'interaction (201) ; déterminer, au moyen du calcul de la corrélation entre les données de comportement d'interaction et un ensemble de données pré-collectées, un premier résultat de sexe réfléchi par les données de comportement d'interaction, l'ensemble de données contenant des données de comportement d'interaction historiques collectées pour un utilisateur, dont le sexe est déterminé ; et obtenir, au moyen de la réalisation d'une analyse de sexe sur les données de temps d'interaction, un second résultat de sexe réfléchi par les données de temps d'interaction (203) ; et déterminer, selon le premier résultat de sexe et le second résultat de sexe, le sexe d'un utilisateur effectuant une interaction d'utilisateur chez le client (205). Selon le procédé, le sexe d'un utilisateur est déterminé sur la base d'un comportement d'interaction entre l'utilisateur et un client, sans qu'il soit nécessaire d'effectuer une reconnaissance faciale sur l'utilisateur.
PCT/CN2021/070935 2020-03-02 2021-01-08 Appareil et procédé d'identification de sexe, dispositif électronique et support de stockage WO2021175010A1 (fr)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202010138805.8 2020-03-02
CN202010138805.8A CN111488519A (zh) 2020-03-02 2020-03-02 用户性别识别的方法、装置、电子设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021175010A1 true WO2021175010A1 (fr) 2021-09-10

Family

ID=71797685

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/070935 WO2021175010A1 (fr) 2020-03-02 2021-01-08 Appareil et procédé d'identification de sexe, dispositif électronique et support de stockage

Country Status (2)

Country Link
CN (1) CN111488519A (fr)
WO (1) WO2021175010A1 (fr)

Families Citing this family (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111488519A (zh) * 2020-03-02 2020-08-04 深圳壹账通智能科技有限公司 用户性别识别的方法、装置、电子设备及存储介质
CN113486946A (zh) * 2021-07-01 2021-10-08 有米科技股份有限公司 基于图像数据的服装商品性别分类方法及装置

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103530540A (zh) * 2013-09-27 2014-01-22 西安交通大学 基于人机交互行为特征的用户身份属性检测方法
CN106656943A (zh) * 2015-11-03 2017-05-10 秒针信息技术有限公司 一种网络用户属性的匹配方法及装置
CN110020155A (zh) * 2017-12-06 2019-07-16 广东欧珀移动通信有限公司 用户性别识别方法及装置
CN111488519A (zh) * 2020-03-02 2020-08-04 深圳壹账通智能科技有限公司 用户性别识别的方法、装置、电子设备及存储介质

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110569347A (zh) * 2019-09-10 2019-12-13 出门问问信息科技有限公司 一种数据处理方法、装置、存储介质和电子设备

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103530540A (zh) * 2013-09-27 2014-01-22 西安交通大学 基于人机交互行为特征的用户身份属性检测方法
CN106656943A (zh) * 2015-11-03 2017-05-10 秒针信息技术有限公司 一种网络用户属性的匹配方法及装置
CN110020155A (zh) * 2017-12-06 2019-07-16 广东欧珀移动通信有限公司 用户性别识别方法及装置
CN111488519A (zh) * 2020-03-02 2020-08-04 深圳壹账通智能科技有限公司 用户性别识别的方法、装置、电子设备及存储介质

Also Published As

Publication number Publication date
CN111488519A (zh) 2020-08-04

Similar Documents

Publication Publication Date Title
US20230013306A1 (en) Sensitive Data Classification
US11501514B2 (en) Universal object recognition
Li et al. Unobservable re-authentication for smartphones.
CN105630856B (zh) 在线用户简档的自动聚合
US11704500B2 (en) Techniques to add smart device information to machine learning for increased context
WO2019196534A1 (fr) Procédé et appareil de reconnaissance homme/machine basée sur un code de vérification
WO2022105118A1 (fr) Procédé et appareil d'identification d'état de santé basés sur une image, dispositif et support de stockage
WO2021175010A1 (fr) Appareil et procédé d'identification de sexe, dispositif électronique et support de stockage
US20160275302A1 (en) Securing a device using graphical analysis
WO2024060951A1 (fr) Procédé et appareil d'entretien pour services
US10997609B1 (en) Biometric based user identity verification
Lee et al. Confident privacy decision-making in IoT environments
US20160275278A1 (en) Securing a device using graphical analysis
Neal et al. You are not acting like yourself: A study on soft biometric classification, person identification, and mobile device use
Alkhamees et al. Event detection from social network streams using frequent pattern mining with dynamic support values
CN115271931A (zh) 一种信用卡产品的推荐方法、装置、电子设备和介质
CN110162535B (zh) 用于执行个性化的搜索方法、装置、设备以及存储介质
Tsimperidis et al. User attribution through keystroke dynamics-based author age estimation
CN112041847A (zh) 提供具有隐私标签的图像
Kwon et al. HaaS (Human activity analytics as a Service) using sensor data of smart devices
CN113642519A (zh) 一种人脸识别系统和人脸识别方法
dos Santos et al. Trends in user identity and continuous authentication
US20240098086A1 (en) Systems and methods for determining trusted devices
KR102619522B1 (ko) 인공지능 기반의 기밀정보 유출 탐지 방법 및 장치
KR102620689B1 (ko) 도어 영상 분석 기반의 ai 비서 장치 및 시스템

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21763724

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 120123)

122 Ep: pct application non-entry in european phase

Ref document number: 21763724

Country of ref document: EP

Kind code of ref document: A1