For example, certain words are used in the description and claims to refer to specific components. Those skilled in the art should understand that hardware manufacturers may use different terms to refer to the same component. This specification and claims do not use differences in names as a way to distinguish components, but use differences in functions of components as a criterion. If the "include" mentioned in the entire specification and claims is an open term, it should be interpreted as "include but not limited to". "Approximately" means that within the acceptable error range, those skilled in the art can solve the technical problem within a certain error range, and basically achieve the technical effect. In addition, the term "coupled" herein includes any direct and indirect electrical coupling means. Therefore, if it is described that a first device is coupled to a second device, it means that the first device can be directly electrically coupled to the second device, or indirectly electrically coupled through other devices or coupling means. Connected to the second device. The subsequent description of the specification is a preferred embodiment for implementing the application, but the description is for the purpose of explaining the general principles of the application and is not intended to limit the scope of the application. The protection scope of this application shall be subject to those defined by the appended claims.
 The cross-analysis based on user behavior is an extremely open analysis method. Before analyzing user data, it is not even necessary to set an expected result. In fact, using data to analyze user behavior is to try to discover some intuitively imperceptible behavior patterns of users from a statistical point of view, so as to further infer the psychological needs of users. Finally, the user's psychological needs are reflected in the product design. Therefore, the design and improvement of products have a clear pertinence to the psychological needs of users. It is an effective means to improve product user experience.
 In this application, it is in accordance with this process to conduct a deeper and more comprehensive analysis of user data. That is to set up multiple dimensions for users, and collect user behavior data on a large scale; and then perform various open analysis operations on large-scale user data on various dimensions. For details, refer to the following embodiments.
 See figure 1 Shown are specific embodiments of the method described in this application. The method in this embodiment specifically includes the following steps:
 Step 101: Establish a user database. Several dimensions are set in the user database, and each dimension corresponds to one or more tag names.
 In this embodiment, unlike the traditional sampling questionnaire survey, a special user database will be established first to collect and count various data related to user behavior. And in the database, set a number of user dimensions describing the user. For example, information such as the user's age, gender, occupation, tendency, activity, and operation records are all dimensions that describe the user.
 And there are several label names corresponding to each dimension, which represent the attributes of the user in that dimension. For example, in this embodiment, age, gender, and occupation are all dimensions describing the characteristics of the user. In the age dimension, the corresponding tag names are <0~15 years old> , <15~25 years old> , <25~35 years old> , <35~50 years old> with <50 years old and above> , Each user must have a tag name that matches his own situation in this dimension. Similarly, the label names corresponding to the gender dimension include with; The label names corresponding to the occupational dimension include , , , or and many more. Dimensions such as tendency, activity, and operation records describe the specific situation of users using network services. The orientation dimension can measure the user’s interest orientation, including tag names such as , or Wait. The activeness dimension can measure how frequently users participate in network services. The operation record dimension specifically records the user's browsing content.
 Since the analysis method described in this embodiment is extremely open, the larger the user information sample saved and the more comprehensive the coverage, the more abundant the results of the analysis will be. Therefore, in the actual user database, the dimensions included and the tag names corresponding to the dimensions are extremely diverse, so I won't list more here.
 Step 102: Collect and save user data of multiple users in the user data collection of the user database. The user data of each user includes the user ID of the user and the matching tag name of the user in one or more dimensions.
 The user data set takes an individual user as a basic unit, and includes multiple or large amounts of user data. Each piece of user data describes various related attributes of a user. Specifically, the user data includes the user ID of the user and the matching tag name of the user in one or more dimensions.
 For example, for user A, its user data can be expressed as follows:
 . Among them, "User A_0001" is the ID of the user, and the rest are the label names of the user in a certain dimension.
 According to the above model, each user has a similar user data. Therefore, the user database can collect user data of a large number or even all users and save it in the user data collection.
 Step 103: Collect and save tag data corresponding to each dimension in the tag set of the user database. Each tag data includes a tag name and a number of user business cards matching the tag.
 There is also a tag set in the user database. The tag data is stored in the tag set as the basic unit. There must be a certain number of user groups in the user database that all have the same label name, then the label name and the user group constitute a label data. That is, a label data will include a label name and user business cards of users in the user group. The user name card may include three types of information: user ID, user weight index, and user support index. The weight index means the user's influence in the user group; the support index means the degree to which the user matches the tag name.
 For example, label The label data is as follows:
 ; j;Y j; k;Y k; l;Y l.
 Where x j , X k , X l Is the weight index, y j , Y k , Y l To support the index, both are numbers.
 In fact, there may be thousands of user groups for a tag, and only the form of the tag data is simply displayed here.
 According to the above mode, each label name has a similar label data. Therefore, the user database can collect a large amount or even all the label data of the label name, and save it in the label data collection.
 In this embodiment, through step 101 to step 103, the establishment of the user database is realized, and sufficient data resources are stored in the user data collection and the tag data collection for the analysis of user behavior.
 In addition, preferably, in order to facilitate the maintenance of the user data collection and the tag collection, in this embodiment, a user list may also be set, in which the user IDs of all users are stored; and a tag list is set, in which all the tags are stored. Label name. Directly manage all user IDs and tag names through the above two lists.
 Step 104: Preset a cross analysis calculation strategy, call the user data set and/or tag set according to the cross analysis calculation strategy, and extract the user data and/or tag data in the user data set and/or tag set.
 From step 104, the data resources stored above, that is, user data and tag data, are used to analyze user behavior.
 It is known from the foregoing that the user behavior analysis described in this embodiment is an open analysis, and the purpose is to try to discover certain intuitively imperceptible behavior patterns of the user from a statistical point of view. It is not even necessary to set an expected analysis result. Therefore, the corresponding analysis process, that is, the content of the cross-analysis calculation strategy should be very diverse and can be set and adjusted arbitrarily.
 Based on the above, in this embodiment, the cross-analysis calculation strategy is not specifically limited. All relevant statistical calculation and analysis methods can be combined under the overall technical solution of this embodiment. Generally speaking, the analysis operations included in the cross-analysis calculation strategy mainly include several important operations such as projection classification operations, screening operations, and/or set intersection operations, used alone or in combination with each other.
 Step 105: Analyze and calculate the extracted user data and/or tag data by using the cross-analysis calculation strategy to obtain user behavior data.
 By using the cross-analysis calculation strategy to calculate and analyze the user data/tag data in the user database, some potential laws in user behavior can be discovered. In this embodiment, this rule can be used as user behavior data.
 The user behavior data often shows intuitive and imperceptible user behavior characteristics, which can help infer the deeper psychological activities and needs of various users when using network services; and in the process of product design and optimization , Make effective use of users’ psychological activities and needs to achieve the purpose of improving user experience.
 From the above technical solutions, it can be seen that the beneficial effects of this embodiment are: establishing a user database and collecting a large amount of data, so that user behavior analysis has a larger scale of data samples, making the analysis results more accurate and rich; by setting diversified cross-analysis The calculation strategy realizes open user behavior analysis, which can obtain the user's potential behavior rules more widely, so as to better understand the user's psychological activities and usage needs in the follow-up, and improve the user experience.
 See figure 2 As shown, this is another specific embodiment of the method described in this application. In this example, figure 1 Based on the illustrated embodiment, a cross-analysis calculation strategy is described in more detail, so as to make the overall technical solution of the present application easier to understand. The data collection stage of the overall technical solution in this embodiment is consistent with the foregoing steps 101 to 103, and the description is not repeated here. Only describe the specific process of using cross-analysis calculation strategy to analyze user behavior.
 Step 201: Extract user data in the user data set, and use a filtering operation to find the The tagged users get the user group α.
 In this embodiment, the first step of user behavior analysis is to screen out a specific user group. Specifically, some users who are interested in the variety show "Where is Dad" are screened out. It can be found in the user database, It belongs to a label in the orientation dimension, so in this embodiment, all user data with this label is found from the user data set, as the user group α, that is, the user group interested in "Where is Dad" is obtained.
 Step 202: Use a screening algorithm to find out that there are tags in the user group α Of users, get the user group β.
 After the user group α is obtained, the user group α is further screened by the geographical dimension to obtain the Beijing users in the user group α. In the user database, Is a label in the geographic dimension, so the user group α is found with Tag users get the Beijing user group who pays attention to "Where is Dad?", and this user group is the user group β.
 Step 203: Use the screening algorithm to find out that there are tags in the user group β Of users, get user group γ.
 In the same way as the foregoing, the user group that uses iphone to receive network services is further screened out in the user group β as the user group γ.
 That is to say, the users in the user group γ all meet the following characteristics: users in Beijing who use iPhone to follow "Where is Dad?"
 Step 204: Calculate the percentage of the user group γ to the user group β, and use the percentage data as user behavior data.
 Based on the analysis results obtained above, the percentage of user group γ to user group β can be further calculated. That is to know what percentage of users in the user group β use their iPhones to follow "Where is Dad?"
 It is assumed that the data result obtained in this embodiment is that the user group γ accounts for 75% of the user group β. This shows that in Beijing, most users who like "Where is Dad" use iPhone, or that users who use iPhone have a higher interest in "Where is Dad". The above is the law of potential usage learned by the user in this embodiment.
 In this embodiment, the following conclusion is obtained through analysis: "75% of users who follow "Where is Dad?" in Beijing use iPhone. This conclusion can be used as the user behavior data obtained in this analysis.
 It can be seen from the above technical solutions that the beneficial effect of this embodiment is that a cross-analysis calculation strategy is specifically described, which makes the overall technical solution of the application more complete and more fully disclosed.
 It can be seen from the above embodiments that the beneficial effects of this application are:
 (1) Establish a user database and collect a large amount of data, so that user behavior analysis has a larger scale of data samples, making the analysis results more accurate and rich;
 (2) By setting up diversified cross-analysis calculation strategies, open user behavior analysis is realized, and users' potential behavior rules can be obtained more widely;
 (2) Through the obtained user behavior data, the user's psychological activities and usage needs can be better understood, and the user experience can be improved.
 Those skilled in the art should understand that the embodiments of the present application may be provided as methods, devices, or computer program products. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
 The above description shows and describes several preferred embodiments of this application, but as mentioned above, it should be understood that this application is not limited to the form disclosed herein, and should not be regarded as an exclusion of other embodiments, but can be used for each Other combinations, modifications, and environments can be modified through the above teachings or technology or knowledge in related fields within the scope of the inventive concept described herein. The modifications and changes made by those skilled in the art do not depart from the spirit and scope of this application, and should fall within the protection scope of the appended claims of this application.