Information identification method and device

Abstract

The embodiments of the present application disclose an information identification method and device. A specific implementation of the method includes: obtaining information associated with a user account to be processed, the information including at least one device identifier and user behavior information corresponding to each device identifier; performing feature extraction on the user behavior information corresponding to each device identifier, inputting the extracted feature information into a pre-trained recognition model associated with a specified user attribute, and obtaining a recognition result, the recognition result including the probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is a preset attribute value; based on the probability in the obtained recognition result, determining whether the attribute value of the specified user attribute of the user to whom the device indicated by at least one device identifier belongs is a preset attribute value. This implementation realizes the identification of the specified user attribute information of the user to whom the devices indicated by different device identifiers associated with the same user account belong.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and specifically to an information identification method and device. Background Art

[0002] On the Internet, users can obtain corresponding user accounts by registering. For some user accounts, the user information associated with the user account may lack some user attribute information, such as gender, age, occupation, etc.

[0003] The existing methods for identifying user attribute information usually identify the user's gender and / or age based on the voice information or video information of the user to whom the user account belongs. However, the same user account may be logged in on different user devices, and the different user devices may belong to different users. Therefore, the same user account may correspond to different shopping habits and shopping preferences on different user devices. Therefore, the user attribute information identified by the existing methods for identifying user attribute information is usually less accurate. Summary of the invention

[0004] The embodiments of the present application provide an information identification method and device.

[0005] In a first aspect, an embodiment of the present application provides an information identification method, the method comprising: obtaining information associated with a user account to be processed, wherein the information comprises at least one device identifier and user behavior information corresponding to the at least one device identifier respectively; performing feature extraction on the user behavior information corresponding to each device identifier, inputting the extracted feature information into a pre-trained recognition model associated with a specified user attribute, and obtaining a recognition result corresponding to the device identifier, wherein the recognition result comprises a probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is a preset attribute value; based on the probability in the obtained recognition result, determining whether the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identifier belongs is the preset attribute value.

[0006] In some embodiments, the above method also includes: generating tag information corresponding to the device identifier in the above at least one device identifier, wherein the above tag information is used to indicate whether the attribute value of the above specified user attribute of the user to whom the device indicated by the corresponding device identifier belongs is the above preset attribute value.

[0007] In some embodiments, the above-mentioned determination of whether the attribute value of the above-mentioned designated user attribute of the user to whom the device indicated by the above-mentioned at least one device identifier belongs is the above-mentioned preset attribute value based on the probability in the obtained recognition result includes: for each device identifier among the above-mentioned at least one device identifier, determining whether the probability in the recognition result corresponding to the device identifier is less than the probability threshold, and if not, determining that the attribute value of the above-mentioned designated user attribute of the user to whom the device indicated by the device identifier belongs is the above-mentioned preset attribute value.

[0008] In some embodiments, the above-mentioned determination of whether the attribute value of the above-mentioned designated user attribute of the user to whom the device indicated by the above-mentioned at least one device identifier belongs is the above-mentioned preset attribute value based on the probability in the obtained recognition result also includes: sorting the probabilities in each obtained recognition result in descending order of numerical value; starting from the maximum value side of the sorted probabilities, selecting a preset number of probabilities that are not less than the probability threshold, and determining the attribute value of the above-mentioned designated user attribute of the user to whom the device indicated by the device identifier corresponding to the recognition result where the selected probability belongs as the above-mentioned preset attribute value.

[0009] In some embodiments, user behavior information includes at least one of the following: browsing information, ordering information, the above browsing information and the above ordering information respectively include information sequence groups, and the information in the information sequence is one of the following: item identification, category identification, brand identification, store identification.

[0010] In some embodiments, the above-mentioned feature extraction of user behavior information corresponding to each device identifier includes: for each information sequence in the user behavior information corresponding to each device identifier, mapping each piece of information in the information sequence into a vector with a specified dimension, determining the average value of each vector obtained by mapping, and using the above-mentioned average value as the extracted feature information.

[0011] In some embodiments, the above-mentioned feature extraction of the user behavior information corresponding to each device identifier also includes: for each information sequence in the user behavior information corresponding to each device identifier, determining the number of information in the information sequence, and using the above number as the extracted feature information.

[0012] In some embodiments, the above-mentioned recognition model is obtained by training through the following training steps: obtaining a sample set, wherein the sample includes user behavior information corresponding to a device identifier and a data tag used to indicate whether the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identifier belongs is the above-mentioned preset attribute value; using a machine learning method to train a recognition model based on the user behavior information and data tags included in the samples in the above-mentioned sample set.

[0013] In some embodiments, the above-mentioned acquisition of the sample set includes: acquiring a preset user account set, wherein at least one user account corresponds to the above-mentioned designated user attribute with an attribute value that is not empty, and each user account also corresponds to at least one device identifier, and the device identifier corresponds to user behavior information; selecting a user account that meets preset conditions from the above-mentioned user account set as the target user account, wherein the above-mentioned preset conditions include: corresponding to a device identifier and the above-mentioned designated user attribute with an attribute value that is not empty; generating the above-mentioned sample set based on the user behavior information corresponding to the device identifier corresponding to the above-mentioned target user account and the attribute value of the above-mentioned designated user attribute corresponding to the above-mentioned target user account.

[0014] In some embodiments, each user account also corresponds to a membership level and a registration time; and the above-mentioned preset conditions also include at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within the first preset time period, the corresponding membership level is not lower than the preset level, and the corresponding registration time is within the second preset time period.

[0015] In a second aspect, an embodiment of the present application provides an information identification device, which includes: an acquisition unit, configured to acquire information associated with a user account to be processed, wherein the information includes at least one device identifier and user behavior information corresponding to the at least one device identifier; an identification unit, configured to extract features of the user behavior information corresponding to each device identifier, and input the extracted feature information into a pre-trained recognition model associated with a specified user attribute to obtain a recognition result corresponding to the device identifier, wherein the recognition result includes a probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is a preset attribute value; a determination unit, configured to determine whether the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identifier belongs is the preset attribute value based on the probability in the obtained recognition result.

[0016] In some embodiments, the above-mentioned device also includes: a generation unit, configured to generate tag information corresponding to the device identifier in the above-mentioned at least one device identifier, wherein the above-mentioned tag information is used to indicate whether the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the corresponding device identifier belongs is the above-mentioned preset attribute value.

[0017] In some embodiments, the above-mentioned determination unit includes: a first determination subunit, configured to determine, for each device identification of the above-mentioned at least one device identification, whether the probability in the identification result corresponding to the device identification is less than a probability threshold; if not, determining that the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identification belongs is the above-mentioned preset attribute value.

[0018] In some embodiments, the above-mentioned determination unit also includes: a sorting subunit, configured to sort the probabilities in each obtained recognition result in descending order of numerical value; a second determination subunit, configured to start from the maximum value side of the sorted probabilities, select a preset number of probabilities that are not less than the probability threshold, and determine the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identifier corresponding to the recognition result where the selected probability belongs as the above-mentioned preset attribute value.

[0019] In some embodiments, user behavior information includes at least one of the following: browsing information, ordering information, the above browsing information and the above ordering information respectively include information sequence groups, and the information in the information sequence is one of the following: item identification, category identification, brand identification, store identification.

[0020] In some embodiments, the above-mentioned identification unit is further configured to: for each information sequence in the user behavior information corresponding to each device identifier, map each piece of information in the information sequence into a vector with a specified dimension, determine the average value of each vector obtained after mapping, and use the above-mentioned average value as the extracted feature information.

[0021] In some embodiments, the identification unit is further configured to: for each information sequence in the user behavior information corresponding to each device identifier, determine the number of information in the information sequence, and use the number as the extracted feature information.

[0022] In some embodiments, the above-mentioned recognition model is obtained by training through the following training steps: obtaining a sample set, wherein the sample includes user behavior information corresponding to a device identifier and a data tag used to indicate whether the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identifier belongs is the above-mentioned preset attribute value; using a machine learning method to train a recognition model based on the user behavior information and data tags included in the samples in the above-mentioned sample set.

[0023] In some embodiments, the above-mentioned acquisition of the sample set includes: acquiring a preset user account set, wherein at least one user account corresponds to the above-mentioned designated user attribute with an attribute value that is not empty, and each user account also corresponds to at least one device identifier, and the device identifier corresponds to user behavior information; selecting a user account that meets preset conditions from the above-mentioned user account set as the target user account, wherein the above-mentioned preset conditions include: corresponding to a device identifier and the above-mentioned designated user attribute with an attribute value that is not empty; generating the above-mentioned sample set based on the user behavior information corresponding to the device identifier corresponding to the above-mentioned target user account and the attribute value of the above-mentioned designated user attribute corresponding to the above-mentioned target user account.

[0024] In some embodiments, each user account also corresponds to a membership level and a registration time; and the above-mentioned preset conditions also include at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within the first preset time period, the corresponding membership level is not lower than the preset level, and the corresponding registration time is within the second preset time period.

[0025] In a third aspect, an embodiment of the present application provides an electronic device, comprising: one or more processors; a storage device for storing one or more programs; when the one or more programs are executed by the one or more processors, the one or more processors implement a method as described in any implementation method in the first aspect.

[0026] In a fourth aspect, an embodiment of the present application provides a computer-readable medium having a computer program stored thereon, and when the program is executed by a processor, the method described in any implementation manner in the first aspect is implemented.

[0027] The information identification method and device provided in the embodiments of the present application obtain information associated with the user account to be processed, including at least one device identifier and the user behavior information corresponding to the at least one device identifier, so as to extract features of the user behavior information corresponding to each device identifier, and input the extracted feature information into a pre-trained recognition model associated with the specified user attribute to obtain a recognition result corresponding to the device identifier. Then, based on the probability in the obtained recognition result, it can be determined whether the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identifier is respectively belongs is a preset attribute value. In this way, the probability of the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identifier corresponding to the user account to be processed is effectively utilized to determine that it is the preset attribute value, and the identification of the specified user attribute information of the user to whom the device indicated by different device identifiers associated with the same user account is respectively belongs is realized. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Other features, objects and advantages of the present application will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings:

[0029] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0030] Figure 2 is a flow chart of an embodiment of an information identification method according to the present application;

[0031] Figure 3 is a schematic diagram of an application scenario of the information identification method according to the present application;

[0032] Figure 4 is a flow chart of an embodiment of a method for generating a sample set according to the present application;

[0033] Figure 5 is a schematic structural diagram of an embodiment of an information identification device according to the present application;

[0034] Figure 6 It is a structural diagram of a computer system suitable for implementing an electronic device of an embodiment of the present application. DETAILED DESCRIPTION

[0035] The present application will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the relevant invention, rather than to limit the invention. It should also be noted that, for ease of description, only the parts related to the relevant invention are shown in the accompanying drawings.

[0036] It should be noted that, in the absence of conflict, the embodiments and features in the embodiments of the present application can be combined with each other. The present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0037] Figure 1 An exemplary system architecture 100 to which an embodiment of the information identification method or information identification device of the present application can be applied is shown.

[0038] like Figure 1 As shown, the system architecture 100 may include an information identification server 101, a network 102, and an information storage server 103. The network 102 is used to provide a medium for a communication link between the information identification server 101 and the information storage server 103. The network 102 may include various connection types, such as wired, wireless communication links, or optical fiber cables, etc.

[0039] The information identification server 101 may be a server that provides various services, such as a server that identifies user attribute information. The information identification server 101 may obtain information associated with the user account to be processed (such as at least one device identifier and user behavior information corresponding to the at least one device identifier), and perform analysis and other processing on the information.

[0040] The information storage server 103 may be a server that provides various services, for example, a server that stores information associated with a user account to be processed.

[0041] It should be noted that the information identification method provided in the embodiment of the present application is generally executed by the information identification server 101, and accordingly, the information identification device is generally arranged in the information identification server 101.

[0042] It should be noted that when the information associated with the user account to be processed is pre-stored locally in the information identification server 101 , the system architecture 100 may not include the information storage server 103 .

[0043] In addition, the information identification server 101 can be hardware or software. When the information identification server 101 is hardware, it can be implemented as a distributed server cluster consisting of multiple servers, or it can be implemented as a single server. When the information identification server 101 is software, it can be implemented as multiple software or software modules (for example, used to provide distributed services), or it can be implemented as a single software or software module. No specific limitation is made here.

[0044] It should be understood that Figure 1 The number of information identification servers, networks and information storage servers in the embodiment is only for illustration. Any number of information identification servers, networks and information storage servers may be provided according to the implementation requirements.

[0045] Continue to refer Figure 2 , shows a process 200 of an embodiment of an information identification method according to the present application. The process 200 of the information identification method includes the following steps:

[0046] Step 201, obtaining information associated with the user account to be processed.

[0047] In this embodiment, the execution subject of the information identification method (for example Figure 1 The information identification server 101 shown in the figure can be connected to the information storage server (for example) by wired connection or wireless connection. Figure 1 The information storage server 103 shown in the figure obtains the information associated with the user account to be processed. Of course, the above execution subject can also obtain the information associated with the user account to be processed locally.

[0048] The user account may be associated with user information, which may include attribute information of the user to which the user account belongs, such as age, gender, occupation, education, date of birth, etc. The user account to be processed may be a user account whose associated user information lacks attribute values ​​of specified user attributes. The information associated with the user account to be processed may include at least one device identifier and user behavior information corresponding to the at least one device identifier.

[0049] It should be noted that the above-mentioned designated user attributes may be user attributes whose attribute values ​​can be predicted by the user behavior information of the user, such as gender, age or purchasing power, etc. It should be understood that the above-mentioned designated user attributes may be set according to actual needs, and this embodiment does not make any limitation to this aspect.

[0050] For any device identifier associated with the user account to be processed, the device indicated by the device identifier may be a device that has logged into the user account to be processed. The user behavior information corresponding to the device identifier may be information generated by browsing, adding items to a shopping cart (i.e., adding item information to a shopping cart), and / or placing an order on the device using the user account to be processed within a specified time period (e.g., the past three months or six months).

[0051] In some optional implementations of this embodiment, the user behavior information may include at least one of the following: browsing information, order information. The browsing information and order information may include information sequence groups respectively. The information in the information sequence may be, for example, one of the following: item identification, category identification, brand identification, store identification. In addition, for any information sequence, the information in the information sequence may be arranged in chronological order of browsing time or ordering time.

[0052] Step 202 , extracting features from the user behavior information corresponding to each device identifier, inputting the extracted feature information into a pre-trained recognition model associated with a specified user attribute, and obtaining a recognition result corresponding to the device identifier.

[0053] In this embodiment, for each of the at least one device identification, the execution subject may extract features of the user behavior information corresponding to the device identification, input the extracted feature information into a pre-trained recognition model associated with the specified user attribute, and obtain a recognition result corresponding to the device identification. The recognition result may include the probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identification belongs is a preset attribute value. The recognition model may be used to characterize the correspondence between the feature information and the recognition result.

[0054] As an example, if the above-mentioned designated user attribute is gender, the above-mentioned preset attribute value may be, for example, female, male, or neutral, etc. If the above-mentioned designated user attribute is age, the above-mentioned preset attribute value may be, for example, a designated age group, such as 18-25 or 26-35, etc. It should be understood that the above-mentioned preset attribute value may be set according to actual needs, and this embodiment does not make any limitation to this aspect.

[0055] Here, the above-mentioned execution subject can, for example, use dimensionality reduction methods such as PCA (Principal Components Analysis) and LDA (Latent Dirichlet Allocation) to reduce the dimensionality of user behavior information, and use the information obtained after dimensionality reduction as feature information. Among them, the PCA method is a statistical method that converts a set of variables that may be correlated into a set of linearly unrelated variables through orthogonal transformation. The converted set of variables is called principal components. LDA is a type of classification algorithm. LDA generally projects historical data to ensure that data of the same category is as close as possible after projection, and data of different categories are as separated as possible, and generates a linear discriminant model to separate and predict newly generated data.

[0056] It should be pointed out that the above-mentioned recognition model can be a correspondence table that is pre-formulated by technicians based on statistics of a large amount of feature information and recognition results, and stores the correspondence between multiple feature information and recognition results. The above-mentioned recognition model can also be obtained by training using a logistic regression model, a naive Bayes model, a random forest model, a gradient progressive regression tree model, etc. Among them, when using the gradient progressive regression tree model, the LightGBM version can be used. LightGBM is a gradient boosting framework that uses a decision tree based on a learning algorithm. It can be said to be distributed and efficient, with the following advantages: faster training efficiency, low memory usage, higher accuracy, support for parallel learning, and the ability to process large-scale data.

[0057] In some optional implementations of this embodiment, when the above-mentioned execution subject is a distributed server cluster composed of multiple servers, if the above-mentioned recognition model is a model obtained by LightGBM training, then the above-mentioned recognition model can be a distributed version of LightGBM directly running on the distributed server cluster. When the above-mentioned execution subject is a single server, if the above-mentioned recognition model is a model obtained by LightGBM training, then the above-mentioned execution subject can use the above-mentioned recognition model in a multi-process parallel manner to perform corresponding prediction operations. In this way, the prediction speed of the recognition model can be effectively improved.

[0058] In some optional implementations of this embodiment, for each device identification in the at least one device identification, for each information sequence in the user behavior information corresponding to the device identification, the execution subject may map each piece of information in the information sequence into a vector with a specified dimension, and then determine the average value of each vector obtained by mapping, and use the average value as the extracted feature information. In addition, the execution subject may also determine the number of information in the information sequence, and use the number as the extracted feature information.

[0059] Here, the execution subject can use a word vector model (such as Word2Vec) to map each piece of information in the information sequence into a vector with a specified dimension. Among them, Word2Vec can train an N-gram model (sometimes also called an N-gram model) based on a neural network, and can also find the word vector corresponding to the word during the training process, which has the effect of reducing the dimension and retaining its semantic information. Compared with traditional dimensionality reduction methods, Word2Vec requires fewer computing resources and is more accurate in representing semantic information.

[0060] In some optional implementations of this embodiment, the above-mentioned recognition model can be obtained by training a model training end (such as the above-mentioned execution subject or a server remotely connected to the above-mentioned execution subject) by executing the following training steps:

[0061] First, the model training end can obtain a sample set. The sample may include user behavior information corresponding to a device identifier and a data tag indicating whether the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is the preset attribute value. Here, the data tag may include, for example, the number 0 or 1. 0 can be used to indicate that the attribute value of the specified user attribute of the user is not the preset attribute value. 1 can be used to indicate that the attribute value of the specified user attribute of the user is the preset attribute value.

[0062] It should be noted that the sample set can be pre-generated or generated by the above model training end. Figure 4 The sample set is generated by the process 400 of the sample set generation method shown, and this embodiment does not make any limitation to this aspect.

[0063] Then, the model training end can use a machine learning method to train a recognition model based on the user behavior information and data tags included in the samples in the sample set. Here, the model training end can divide the sample set into a training sample set and a test sample set. The recognition model is trained using the samples in the training sample set, and the accuracy of the recognition model is predicted using the samples in the test sample set.

[0064] In the training process, for each sample in the training sample set, the model training end may extract features of the user behavior information in the sample, input the extracted feature information into the recognition model in training, and obtain a recognition result. The model training end may use a preset classification loss function to determine the difference between the probability in the recognition result and the data label in the sample, and the model training end may use a preset back propagation algorithm to adjust the parameters in the recognition model according to the difference.

[0065] It should be noted that the above classification loss function can be various loss functions for classification (such as HingeLoss function or Softmax Loss function, etc.). During the training process, the training goal is to minimize the value of the classification loss function. Therefore, the parameters of the recognition model obtained after training are the parameters corresponding to the minimum value of the classification loss function.

[0066] In addition, the above-mentioned back propagation algorithm can also be called an error back propagation algorithm, or an error reverse propagation algorithm. The learning process of the back propagation algorithm consists of a forward propagation process and a backward propagation process. In a feedforward network, the input signal is input through the input layer, calculated by the hidden layer, and output by the output layer. The output value is compared with the label value. If there is an error, the error is propagated backward from the output layer to the input layer. In this process, the gradient descent algorithm (such as the stochastic gradient descent algorithm) can be used to adjust the neuron weights.

[0067] Step 203: Based on the probability in the obtained recognition result, it is determined whether the attribute value of the designated user attribute of the user to which the device indicated by at least one device identifier belongs is a preset attribute value.

[0068] In this embodiment, the execution subject may determine whether the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identifier belongs is the preset attribute value based on the probability in the obtained recognition result.

[0069] As an example, for each of the at least one device identification mentioned above, the above-mentioned execution entity can determine whether the probability in the recognition result corresponding to the device identification is less than a probability threshold (for example, 0.45 or 0.5, etc.); if not, the above-mentioned execution entity can determine that the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identification belongs is the above-mentioned preset attribute value.

[0070] In some optional implementations of this embodiment, the execution subject may first sort the probabilities in the obtained recognition results in descending order of numerical values. Then, the execution subject may select a preset number (e.g., 5, 10, or 20, etc.) of probabilities that are not less than the probability threshold from the maximum value side of the sorted probabilities, and determine the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier corresponding to the recognition result where the selected probability belongs as the preset attribute value.

[0071] It should be noted that the above probability threshold and the above preset number can be adjusted according to actual needs, and this embodiment does not impose any limitation on this aspect.

[0072] It should be noted that, for each device identifier associated with the same user account, by determining the specified attribute information of the user to which the device indicated by each device identifier belongs, the comprehensiveness and accuracy of the determined specified user attribute information can be improved. Moreover, it can help to improve the effectiveness of the specified user attribute information in different application scenarios.

[0073] For example, if the gender of different device identifiers associated with the same user account are all identified, then in the push scenario of women's product information, the women's product information to be pushed can be pushed to the user device to which the user belongs is a female, which can attract the attention of female users to the received women's product information and improve the effectiveness of information push. If the gender of each device identifier associated with the same user account is not identified, the women's product information to be pushed may be pushed to the user device to which the user belongs is a male, which may reduce user satisfaction.

[0074] In some optional implementations of the present embodiment, the execution entity may generate tag information corresponding to the device identification in the at least one device identification. The tag information may be used to indicate whether the attribute value of the specified user attribute of the user to whom the device indicated by the corresponding device identification belongs is the preset attribute value. As an example, assume that the specified user attribute is gender, and the preset attribute value is female. For any device identification, if it is determined that the gender of the user to whom the device indicated by the device identification belongs is female, then the tag information generated for the device identification may include the text "female" and the probability that the user's gender is female; otherwise, the tag information may include the text "non-female" and the probability that the user's gender is not female.

[0075] Continue to see Figure 3 , Figure 3 FIG. 1 is a schematic diagram of an application scenario of the information identification method according to this embodiment. Figure 3 In the application scenario, the specified user attribute is gender, the preset attribute value is female, and the probability threshold is 0.5.

[0076] First, the information identification server 301 can obtain information associated with the user account to be processed from the connected information storage server 302. The information may include device identifiers A1 and A2, user behavior information B1 corresponding to the device identifier A1, and user behavior information B2 corresponding to the device identifier A2, as shown in reference numeral 303.

[0077] Then, the information identification server 301 may perform feature extraction on the user behavior information B1 and B2 respectively to obtain feature information 304 corresponding to the user behavior information B1 and feature information 305 corresponding to the user behavior information B2.

[0078] Afterwards, the information identification server 301 may input the feature information 304 and 305 into the identification model associated with gender, and obtain the identification results corresponding to the feature information 304 and 305. The identification result corresponding to the feature information 304 may include the probability 0.6 that the gender of the user User1 to which the device indicated by the device identifier A1 belongs is female, as shown by reference numeral 306; the identification result corresponding to the feature information 305 may include the probability 0.4 that the gender of the user User2 to which the device indicated by the device identifier A2 belongs is female, as shown by reference numeral 307.

[0079] Finally, the information identification server 301 can compare the probabilities 0.6 and 0.4 with 0.5, respectively, and determine that the probability 0.6 is not less than 0.5, and the probability 0.4 is less than 0.5. Therefore, the information identification server 301 can obtain a determination result that the gender of user User1 is female, as shown in label 308; and obtain a determination result that the gender of user User2 is not female, as shown in label 309.

[0080] The method provided by the above-mentioned embodiment of the present application effectively utilizes the probability of determining that the attribute value of the specified user attribute of the user to which the device indicated by at least one device identifier corresponding to the user account to be processed belongs is a preset attribute value, thereby realizing the identification of the specified user attribute information of the user to which the devices indicated by different device identifiers associated with the same user account belong.

[0081] Further references Figure 4 , which shows a process 400 of an embodiment of a method for generating a sample set. The process 400 of the method for generating a sample set includes the following steps:

[0082] Step 401, obtaining a preset user account set.

[0083] In this embodiment, the execution subject of the sample set generation method (for example Figure 1 The information identification server 101 shown in the figure or a server connected to the information identification server 101 in remote communication) can be used from a local or connected information storage server (e.g. Figure 1 The information storage server 103 shown in the figure obtains a preset user account set. Among them, at least one user account may correspond to a specified user attribute whose attribute value is not empty. Each user account may also correspond to a membership level, registration time and at least one device identifier. The device identifier may correspond to user behavior information.

[0084] Step 402: Select a user account that meets a preset condition from the user account set as a target user account.

[0085] In this embodiment, the execution subject may select a user account that meets the preset conditions from the user account set as the target user account. The preset conditions may include the above-mentioned specified user attribute corresponding to a device identifier and an attribute value that is not empty, and at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within a first preset time period (e.g., the past three months or six months, etc.), the corresponding membership level is not lower than the preset level, and the corresponding registration time is within a second preset time period (e.g., the past two years or three years, etc.).

[0086] The user behavior information may include order information, and the order information may include order time. The execution subject may determine whether the user has placed an order within the first preset time period according to each order time in the order information.

[0087] It should be noted that the first preset time period, the preset level and the second preset time period can be adjusted according to actual needs, and this embodiment does not impose any limitation on this aspect.

[0088] Step 403: Generate a sample set based on the user behavior information corresponding to the device identifier corresponding to the target user account and the attribute value of the specified user attribute corresponding to the target user account.

[0089] In this embodiment, the execution subject may generate a sample set based on the user behavior information corresponding to the device identifier corresponding to the target user account and the attribute value of the specified user attribute corresponding to the target user account.

[0090] Here, for each target user account, the execution subject may first generate a data tag based on the attribute value of the specified user attribute corresponding to the target user account, and then generate a sample using the user behavior information corresponding to the device identifier corresponding to the target user account and the data tag. The data tag may be used to indicate whether the attribute value of the specified user attribute corresponding to the target user account is a preset attribute value. Finally, the execution subject may merge the samples generated for each target user account to generate a sample set.

[0091] The sample set generation method provided by the above-mentioned embodiment of the present application effectively utilizes the above-mentioned preset conditions to screen out the target user accounts, which can ensure that the users to whom the screened target user accounts belong are relatively active and loyal users, and the information they register is relatively more credible. Therefore, the sample set generated based on the user behavior information corresponding to the device identifier corresponding to the screened target user account and the attribute value of the specified user attribute corresponding to the target user account can have a high validity. Applying the sample set to the training of the recognition model can improve the prediction accuracy of the recognition model.

[0092] Further references Figure 5 As an implementation of the methods shown in the above figures, the present application provides an embodiment of an information identification device, and the device embodiment is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0093] like Figure 5 As shown, the information identification device 500 of this embodiment includes: an acquisition unit 501, an identification unit 502 and a determination unit 503. The acquisition unit 501 is configured to acquire information associated with the user account to be processed, wherein the information may include at least one device identification and user behavior information corresponding to the at least one device identification; the identification unit 502 is configured to extract features of the user behavior information corresponding to each device identification, input the extracted feature information into a pre-trained recognition model associated with a specified user attribute, and obtain a recognition result corresponding to the device identification, wherein the recognition result may include the probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identification belongs is a preset attribute value; the determination unit 503 is configured to determine whether the attribute value of the specified user attribute of the user to whom the device indicated by the at least one device identification belongs is the preset attribute value based on the probability in the obtained recognition result.

[0094] In this embodiment, in the information identification device 500, the specific processing of the acquisition unit 501, the identification unit 502 and the determination unit 503 and the technical effects thereof can be referred to in Figure 2 The relevant descriptions of step 201, step 202 and step 203 in the corresponding embodiment are not repeated here.

[0095] In some optional implementations of this embodiment, the above-mentioned device 500 may also include: a generation unit (not shown in the figure), configured to generate tag information corresponding to the device identifier in the above-mentioned at least one device identifier, wherein the above-mentioned tag information can be used to indicate whether the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the corresponding device identifier belongs is the above-mentioned preset attribute value.

[0096] In some optional implementations of this embodiment, the above-mentioned determination unit 503 may include: a first determination subunit (not shown in the figure), configured to determine, for each device identification of the above-mentioned at least one device identification, whether the probability in the identification result corresponding to the device identification is less than a probability threshold, and if not, determine that the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identification belongs is the above-mentioned preset attribute value.

[0097] In some optional implementations of the present embodiment, the determination unit 503 may further include: a sorting subunit (not shown in the figure), configured to sort the probabilities in the obtained recognition results in descending order of numerical value; a second determination subunit (not shown in the figure), configured to select a preset number of probabilities that are not less than a probability threshold starting from the maximum value side of the sorted probabilities, and determine the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier corresponding to the recognition result where the selected probability belongs as the preset attribute value.

[0098] In some optional implementations of this embodiment, user behavior information may include at least one of the following: browsing information, order information, and the above browsing information and the above order information may respectively include information sequence groups, and the information in the information sequence may be one of the following: item identification, category identification, brand identification, and store identification.

[0099] In some optional implementations of the present embodiment, the identification unit 502 may be further configured to: for each information sequence in the user behavior information corresponding to each device identifier, map each piece of information in the information sequence into a vector with a specified dimension, determine the average value of each vector obtained after mapping, and use the average value as the extracted feature information.

[0100] In some optional implementations of this embodiment, the above-mentioned identification unit may be further configured to: for each information sequence in the user behavior information corresponding to each device identifier, determine the number of information in the information sequence, and use the above-mentioned number as the extracted feature information.

[0101] In some optional implementations of the present embodiment, the above-mentioned recognition model may be obtained by training through the following training steps: obtaining a sample set, wherein the sample may include user behavior information corresponding to a device identifier and a data tag for indicating whether the attribute value of the above-mentioned specified user attribute of the user to whom the device indicated by the device identifier belongs is the above-mentioned preset attribute value; using a machine learning method to train a recognition model based on the user behavior information and data tags included in the samples in the above-mentioned sample set.

[0102] In some optional implementations of the present embodiment, the above-mentioned acquisition of the sample set may include: acquiring a preset user account set, wherein at least one user account may correspond to the above-mentioned designated user attribute with an attribute value that is not empty, and each user account may also correspond to at least one device identifier, and the device identifier may correspond to user behavior information; selecting a user account that meets preset conditions from the above-mentioned user account set as the target user account, wherein the above-mentioned preset conditions may include: corresponding to a device identifier and the above-mentioned designated user attribute with an attribute value that is not empty; generating the above-mentioned sample set based on the user behavior information corresponding to the device identifier corresponding to the above-mentioned target user account and the attribute value of the above-mentioned designated user attribute corresponding to the above-mentioned target user account.

[0103] In some optional implementations of this embodiment, each user account may also correspond to a membership level and a registration time; and the above-mentioned preset conditions may also include at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within the first preset time period, the corresponding membership level is not lower than the preset level, and the corresponding registration time is within the second preset time period.

[0104] The device provided by the above-mentioned embodiment of the present application effectively utilizes the probability of determining that the attribute value of the specified user attribute of the user to which the device indicated by at least one device identifier corresponding to the user account to be processed belongs is a preset attribute value, thereby realizing the identification of the specified user attribute information of the user to which the devices indicated by different device identifiers associated with the same user account belong.

[0105] Reference below Figure 6 , which shows a schematic diagram of the structure of a computer system 600 suitable for implementing an electronic device of an embodiment of the present application. Figure 6 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0106] like Figure 6 As shown, the computer system 600 includes a central processing unit (CPU) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage part 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the system 600 are also stored. The CPU 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0107] The following components are connected to the I / O interface 605: an input section 606 including a keyboard, a mouse, etc.; an output section 607 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 608 including a hard disk, etc.; and a communication section 609 including a network interface card such as a LAN card, a modem, etc. The communication section 609 performs communication processing via a network such as the Internet. A drive 610 is also connected to the I / O interface 605 as needed. A removable medium 611, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 610 as needed, so that a computer program read therefrom is installed into the storage section 608 as needed.

[0108] In particular, according to an embodiment of the present disclosure, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network through a communication part 609, and / or installed from a removable medium 611. When the computer program is executed by a central processing unit (CPU) 601, the above-mentioned functions defined in the system of the present application are executed.

[0109] It should be noted that the computer-readable medium shown in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than computer-readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0110] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0111] The units involved in the embodiments described in the present application may be implemented by software or hardware. The units described may also be set in a processor, for example, it may be described as: a processor includes an acquisition unit, an identification unit, and a determination unit. The names of these units do not constitute a limitation on the units themselves in some cases, for example, the acquisition unit may also be described as a "unit for acquiring information associated with the user account to be processed".

[0112] As another aspect, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiment; or it may exist independently and not be assembled into the electronic device. The above computer-readable medium carries one or more programs, and when the above one or more programs are executed by an electronic device, the electronic device includes: obtaining information associated with the user account to be processed, wherein the information may include at least one device identification and user behavior information corresponding to the at least one device identification; performing feature extraction on the user behavior information corresponding to each device identification, inputting the extracted feature information into a pre-trained recognition model associated with the specified user attribute, and obtaining a recognition result corresponding to the device identification, wherein the recognition result may include the probability that the attribute value of the above specified user attribute of the user to whom the device indicated by the device identification belongs is a preset attribute value; based on the probability in the obtained recognition result, determining whether the attribute value of the above specified user attribute of the user to whom the device indicated by the at least one device identification belongs is the preset attribute value.

[0113] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above invention concept. For example, the above features are replaced with the technical features with similar functions disclosed in this application (but not limited to) by each other to form a technical solution.

Claims

1. An information identification method, comprising: Acquire information associated with the user account to be processed, wherein the information includes at least one device identifier and user behavior information corresponding to the at least one device identifier, and the device indicated by the device identifier is a device that has logged in to the user account to be processed; Perform feature extraction on the user behavior information corresponding to each device identifier, input the extracted feature information into a pre-trained recognition model associated with the specified user attribute, and obtain a recognition result corresponding to the device identifier, wherein the recognition result includes a probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is a preset attribute value; Based on the probability in the obtained recognition result, it is determined whether the attribute value of the designated user attribute of the user to which the device respectively indicated by the at least one device identification belongs is the preset attribute value.

2. The method according to claim 1, wherein: The method further comprises: Generate tag information corresponding to the device identification in the at least one device identification, wherein the tag information is used to indicate whether the attribute value of the specified user attribute of the user to whom the device indicated by the corresponding device identification belongs is the preset attribute value.

3. The method according to claim 1, wherein: The determining, based on the probability in the obtained recognition result, whether the attribute value of the specified user attribute of the user to which the device indicated by the at least one device identifier belongs is the preset attribute value includes: For each device identification in the at least one device identification, determine whether the probability in the identification result corresponding to the device identification is less than the probability threshold; if not, determine that the attribute value of the specified user attribute of the user to whom the device indicated by the device identification belongs is the preset attribute value.

4. The method according to claim 1, wherein: The determining, based on the probability in the obtained recognition result, whether the attribute value of the specified user attribute of the user to which the device indicated by the at least one device identifier belongs is the preset attribute value, further includes: Sort the probabilities of the obtained recognition results in descending order; Starting from the maximum value side of the sorted probabilities, a preset number of probabilities that are not less than the probability threshold are selected, and the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier corresponding to the recognition result where the selected probability belongs is determined as the preset attribute value.

5. The method according to claim 1, wherein: The user behavior information includes at least one of the following: browsing information and ordering information. The browsing information and the ordering information respectively include an information sequence group. The information in the information sequence is one of the following: item identification, category identification, brand identification, and store identification.

6. The method according to claim 5, wherein: The feature extraction of the user behavior information corresponding to each device identifier includes: For each information sequence in the user behavior information corresponding to each device identifier, each piece of information in the information sequence is mapped into a vector with a specified dimension, and an average value of each vector obtained by mapping is determined, and the average value is used as the extracted feature information.

7. The method according to claim 6, wherein: The feature extraction of the user behavior information corresponding to each device identifier also includes: For each information sequence in the user behavior information corresponding to each device identifier, the number of information in the information sequence is determined, and the number is used as the extracted feature information.

8. The method according to claim 1, wherein: The recognition model is trained through the following training steps: Acquire a sample set, wherein the sample includes user behavior information corresponding to a device identifier and a data tag for indicating whether an attribute value of the specified user attribute of a user to whom the device indicated by the device identifier belongs is the preset attribute value; The recognition model is obtained by training using a machine learning method based on the user behavior information and data labels included in the samples in the sample set.

9. The method according to claim 8, wherein: The obtaining of the sample set comprises: Obtain a preset user account set, wherein at least one user account corresponds to the specified user attribute whose attribute value is not empty, and each user account also corresponds to at least one device identifier, and the device identifier corresponds to user behavior information; Selecting a user account that meets a preset condition from the user account set as a target user account, wherein the preset condition includes: corresponding to a device identifier and the specified user attribute having an attribute value that is not empty; The sample set is generated based on the user behavior information corresponding to the device identification corresponding to the target user account and the attribute value of the designated user attribute corresponding to the target user account.

10. The method according to claim 9, wherein: Each user account also corresponds to a membership level and registration time; and The preset conditions also include at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within the first preset time period, the corresponding membership level is not lower than the preset level, and the corresponding registration time is within the second preset time period.

11. An information recognition device, comprising: an acquisition unit configured to acquire information associated with the user account to be processed, wherein the information includes at least one device identifier and user behavior information corresponding to the at least one device identifier, and the device indicated by the device identifier is a device that has logged into the user account to be processed; an identification unit configured to extract features from the user behavior information corresponding to each device identifier, input the extracted feature information into a pre-trained identification model associated with the specified user attribute, and obtain an identification result corresponding to the device identifier, wherein the identification result includes a probability that the attribute value of the specified user attribute of the user to whom the device indicated by the device identifier belongs is a preset attribute value; The determination unit is configured to determine whether the attribute value of the designated user attribute of the user to which the device respectively indicated by the at least one device identifier belongs is the preset attribute value based on the probability in the obtained recognition result.

12. The device according to claim 11, wherein The device also includes: A generating unit is configured to generate tag information corresponding to a device identifier in the at least one device identifier, wherein the tag information is used to indicate whether the attribute value of the specified user attribute of the user to whom the device indicated by the corresponding device identifier belongs is the preset attribute value.

13. The device according to claim 11, wherein: The user behavior information includes at least one of the following: browsing information and ordering information. The browsing information and the ordering information respectively include an information sequence group. The information in the information sequence is one of the following: item identification, category identification, brand identification, and store identification.

14. The device according to claim 13, wherein: The identification unit is further configured to: For each information sequence in the user behavior information corresponding to each device identifier, each piece of information in the information sequence is mapped into a vector with a specified dimension, and an average value of each vector obtained by mapping is determined, and the average value is used as the extracted feature information.

15. The device according to claim 14, wherein: The identification unit is further configured to: For each information sequence in the user behavior information corresponding to each device identifier, the number of information in the information sequence is determined, and the number is used as the extracted feature information.

16. The device according to claim 11, wherein The recognition model is trained through the following training steps: Acquire a sample set, wherein the sample includes user behavior information corresponding to a device identifier and a data tag for indicating whether an attribute value of the specified user attribute of a user to whom the device indicated by the device identifier belongs is the preset attribute value; The recognition model is obtained by training using a machine learning method based on the user behavior information and data labels included in the samples in the sample set.

17. The device according to claim 16, wherein: The obtaining of the sample set comprises: Obtain a preset user account set, wherein at least one user account corresponds to the specified user attribute whose attribute value is not empty, and each user account also corresponds to at least one device identifier, and the device identifier corresponds to user behavior information; Selecting a user account that meets a preset condition from the user account set as a target user account, wherein the preset condition includes: corresponding to a device identifier and the specified user attribute having an attribute value that is not empty; The sample set is generated based on the user behavior information corresponding to the device identification corresponding to the target user account and the attribute value of the designated user attribute corresponding to the target user account.

18. The device according to claim 17, wherein: Each user account also corresponds to a membership level and registration time; and The preset conditions also include at least one of the following: the user to whom the device indicated by the corresponding device identifier belongs has placed an order at least once within the first preset time period, the corresponding membership level is not lower than the preset level, and the corresponding registration time is within the second preset time period.

19. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 10.

20. A computer readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 10 is implemented.

Images