Information processing method, device and electronic equipment

By using language models in data mining to summarize and summarize the individual attribute values ​​of the target group and generate semantic description information, the problems of low interpretability and low efficiency caused by artificial dependence in the existing technology are solved, and efficient groups are achieved Attribute information acquisition.

CN118734197BActive Publication Date: 2024-11-26UCWEB
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411231672.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-03
Publication Date
2024-11-26
Estimated Expiration
2044-09-03

AI Technical Summary

Technical Problem

Existing data mining technologies rely on manual experience, resulting in a bias in semantic understanding of feature labels. The results are not interpretable, and manual mining consumes huge labor and is inefficient.

Method used

By obtaining the individual attribute values ​​of the target group and entering a preset language model, the language model is used to summarize and summarize, and group attribute information is generated, including individual attribute values ​​and semantic description information, improving interpretability and reducing manual participation.

Benefits of technology

It improves the interpretability and processing efficiency of group attribute information, reduces the need for manual participation, and improves the accuracy and efficiency of attribute mining.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118734197B_ABST
    Figure CN118734197B_ABST
Patent Text Reader

Abstract

The embodiments of the present application provide an information processing method, device and electronic device. The information processing method includes: for a target group, obtaining individual attribute values ​​of individuals included in the target group under a preset individual attribute dimension; inputting each individual attribute value into a preset language model, and obtaining group attribute information for the target group through the preset language model; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute. In the embodiments of the present application, each individual attribute value is summarized and summarized through a language model, so that individual attribute values ​​with similar semantics are classified as the same group attribute, and semantic description information of the group attribute is provided, so that the user can better understand the classified group attributes, that is, the embodiments of the present application can effectively improve the interpretability of the obtained group attribute information. In addition, since the above process does not require human participation, the processing efficiency is relatively high.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computer technology, and in particular, to an information processing method, device and electronic device. Background Art

[0002] Attribute mining, also known as feature mining, refers to the process of searching for hidden features (attributes) from a large amount of group data through algorithms. For example, for a target user group, data mining technology can be used to summarize and generalize the features of multiple pre-set dimensions to analyze the significant features of this user group that are different from other groups.

[0003] For example, for the target group that has watched a short video, we can summarize and generalize the characteristics of different dimensions such as gender, age, occupation, interest preferences, etc., so as to obtain the significant characteristics of the target group. For example, the significant characteristics of the target group are: post-90s female group.

[0004] In the past, data mining mainly relied on manual experience to combine feature labels to mine the significant features of a group. Since people often have deviations in their semantic understanding of feature labels and are limited by manual experience, the results obtained through the above methods are often not interpretable, and manual mining consumes a huge amount of manpower. Summary of the invention

[0005] In view of this, embodiments of the present application provide an information processing method, device, and electronic device to at least partially solve the above problems.

[0006] According to a first aspect of an embodiment of the present application, there is provided an information processing method, including:

[0007] For a target group, obtaining individual attribute values ​​of individuals in the target group under a preset individual attribute dimension;

[0008] Each individual attribute value is input into a preset language model, and group attribute information for the target group is obtained through the preset language model; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute.

[0009] According to a second aspect of an embodiment of the present application, there is provided an information processing device, including:

[0010] An acquisition module, used for acquiring individual attribute values ​​of individuals in a target group under a preset individual attribute dimension;

[0011] The attribute mining module is used to input each individual attribute value into a preset language model, and obtain group attribute information for the target group through the preset language model; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute.

[0012] According to a third aspect of an embodiment of the present application, there is provided an electronic device, comprising: a processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; the memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform an operation corresponding to the method described in the first aspect.

[0013] According to a fourth aspect of an embodiment of the present application, a computer storage medium is provided, on which a computer program is stored. When the program is executed by a processor, the method described in the first aspect is implemented.

[0014] The information processing scheme provided in the embodiment of the present application, after obtaining the individual attribute values ​​of each individual in the target group to be mined under different individual attribute dimensions, inputs each individual attribute value into the language model, and summarizes and summarizes each individual attribute value through the language model, so that the individual attribute values ​​with similar semantics are classified into the same category (that is, classified into the same group attribute), and provides an explanation of the classification (that is, the semantic description information of the group attribute), so that the user can better understand the classified group attributes, that is, the embodiment of the present application can effectively improve the interpretability of the obtained group attribute information. In addition, since the above process does not require human participation, the processing efficiency is relatively high. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of the present application. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0016] Figure 1 A flowchart of a method for processing information according to Embodiment 1 of the present application;

[0017] Figure 2 A schematic diagram of group attributes obtained according to the information processing method of an embodiment of the present application;

[0018] Figure 3 A schematic diagram of constructing a decision tree model based on group attributes obtained in an embodiment of the present application;

[0019] Figure 4A schematic diagram of an information processing flow according to an embodiment of the present application;

[0020] Figure 5 is a structural block diagram of an information processing device according to Embodiment 2 of the present application;

[0021] Figure 6 This is a schematic diagram of the structure of an electronic device according to the third embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable those skilled in the art to better understand the technical solutions in the embodiments of the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the embodiments of the present application should fall within the scope of protection of the embodiments of the present application.

[0023] The specific implementation of the embodiment of the present application is further explained below in conjunction with the accompanying drawings of the embodiment of the present application.

[0024] Embodiment 1

[0025] Reference Figure 1 , Figure 1 The following is a flowchart of an information processing method according to Embodiment 1 of the present application. Specifically, the information processing method provided in this embodiment may include the following steps:

[0026] Step 102: for the target group, obtain individual attribute values ​​of individuals included in the target group under a preset individual attribute dimension.

[0027] Specifically, the target group can also be referred to as the group to be mined, and the group can be a group composed of multiple individuals who meet the preset mining conditions. In the embodiment of the present application, the specific content of the preset mining conditions is not limited, and can be set according to the actual mining task. For example: for an application that provides a certain service to users, the mining task can be: to find out what kind of people are mainly using the application, and what characteristics the people who use the application usually have (also known as: what attributes). In this scenario, the preset mining condition can be that the application has been installed or used, and correspondingly, the target group can be a user group composed of users who have installed or used the application. Through the information processing solution provided in the embodiment of the present application, the product and operation personnel of the application can clearly understand the attributes or characteristics of its target population and target individuals.

[0028] Before conducting group attribute mining, a variety of different individual attribute dimensions (individual feature dimensions) can usually be pre-set for the individuals in the target group to be mined, that is, a variety of different dimensional labels are set for the individuals. The mining process is to semantically summarize and classify the individual attribute values ​​corresponding to the above-mentioned different individual attribute dimensions, so as to obtain the group attributes (group characteristics) possessed by the target group. In the embodiment of the present application, there is no limitation on the specific content of the individual target dimension, as well as the number of individual target dimensions, which can be customized according to actual scenarios and experience. For example: for a target group composed of users, the individual target dimension can be set from one or more of the following aspects: basic characteristics of users, such as age, gender, etc.; user interest preferences for different things, such as interest preferences for film and television series, etc.

[0029] In order to improve the accuracy of attribute mining, in the embodiment of the present application, the obtained individual attribute value (individual feature value) can be represented by natural language with certain semantics. For example, in terms of the individual attribute dimension of gender, for female individuals, their individual attribute value can be: "gender = female", or "gender is female", etc.

[0030] Step 104, input each individual attribute value into a preset language model, and obtain group attribute information for the target group through the preset language model; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute.

[0031] Specifically, in the embodiment of the present application, the group attributes are mined through the pre-trained language model, so as to output the group attribute information (group characteristic information) of the target group. In the embodiment of the present application, the type of the preset language model and the specific results are not limited. For example, the preset language model can be any suitable generative language model.

[0032] When the preset language model is a pre-trained generative language model, a model prompt word can be constructed according to the individual attribute values ​​obtained in step 102, and then the prompt word can be input into the model to trigger the model to perform semantic induction and classification from each individual attribute value, thereby obtaining relevant information about the group attributes possessed by the target group. The embodiment of the present application does not limit the specific form and content of the prompt word, and any prompt word that contains the individual attribute value obtained in step 102 and can enable the model to output the group attribute information of the target group can be used. For example, the prompt word can include: individual attribute value information, and the content of the group attribute information that the model is required to output, etc. For example, the prompt word may be: "I will give you a list of individual attribute values ​​of a certain group of people, where the first value of each item is the attribute variable name, and the second value is the attribute value. The data is: xxxxxx. You need to use the above attribute values ​​to summarize the basic situation of the group of people. Output: group name, group description. The format is as follows: {\"group name\":\"xxx group\",\"group description\":\"xxx\"}. Furthermore, in order to facilitate subsequent processing of the group attributes output by the model, the data format of the group attribute information output by the model can also be set in the prompt word, such as: JSON format, etc.

[0033] The following is an explanation of the process of attribute mining using the preset language model by way of specific examples. Assuming that two different individual attribute values ​​of the individual attribute dimension of age group are obtained through step 102: ["age group", "[30,34]"], and ["age group", "[35,39]", then a model prompt word can be generated based on the above individual attribute values: "I will give you a list of individual attribute values ​​of a certain group of people, where the first value of each item is the attribute variable name and the second value is the attribute value. The data is: ["age group", "[30,34]"], and ["age group", "[35,39]". You need to use the above attribute values ​​to summarize the basic situation of the group. Output: group name, group description. The format is as follows: {\"group name\":\"xxx group\",\"group description\":\"xxx\"}. After the above prompt words are input into the preset language model, the output of the model can be obtained as follows: {"group name":"30-39 years old group","group description":"mainly composed of individuals aged 30-39"}.

[0034] Optionally, in some of the embodiments, the group attribute information may further include: identification information of the group attribute; the group attribute information is generated based on semantic information of individual attribute values ​​contained in the group attribute.

[0035] See also Figure 2 , Figure 2Schematic diagram of group attributes obtained according to the information processing method of the embodiment of the present application. Figure 2 The information processing scheme provided in the embodiment of the present application is explained as follows:

[0036] Assume that in a target application promotion scenario, the individual attribute values ​​of the group using the target application in terms of gender, age, education, and film and television drama interest preferences are obtained (e.g. Figure 2 As shown on the right side, the above individual attribute values ​​can be input into a preset language model, and the model can be used to classify, summarize and generalize the input individual attribute values, thereby outputting group attributes for the above group (see Figure 2 As shown on the left), as well as the individual attribute values ​​contained in each group attribute. Through the group attributes output by the model, the target application operator can clearly identify the target audience of the target application as: female groups born in the 1980s who like to watch realistic emotional movies and TV series.

[0037] The information processing scheme provided in the embodiment of the present application, after obtaining the individual attribute values ​​of each individual in the target group to be mined under different individual attribute dimensions, inputs each individual attribute value into the language model, and summarizes and summarizes each individual attribute value through the language model, so that the individual attribute values ​​with similar semantics are classified into the same category (that is, classified into the same group attribute), and provides an explanation of the classification (that is, the semantic description information of the group attribute), so that the user can better understand the classified group attributes, that is, the embodiment of the present application can effectively improve the interpretability of the obtained group attribute information. In addition, since the above process does not require human participation, the processing efficiency is relatively high.

[0038] In addition, when relying on manual attribute mining, some individual attribute values ​​that have a great impact on downstream prediction tasks but have a low coverage rate in the target group are often ignored during the mining process, which is a loss for the accuracy of the final mining results. In the embodiment of the present application, for individual attribute values ​​with a small coverage rate, they can still be summarized into group attributes with a higher coverage rate through a preset language model. Compared with the manual mining method, the group attributes obtained in the embodiment of the present application can contain more incremental information. Therefore, subsequent model prediction based on the group attributes can effectively improve the accuracy of the prediction results.

[0039] The information processing method of this embodiment can be executed by any appropriate electronic device with data processing capability, including but not limited to: a server, a PC, etc.

[0040] Optionally, in some embodiments, after the above step 102, the method further includes:

[0041] Filter attribute values ​​from each individual attribute value to obtain filtered attribute values;

[0042] Input each individual attribute value into the preset language model, and obtain group attribute information for the target group through the preset language model, including:

[0043] The filtered attribute values ​​are input into a preset language model, and group attribute information for the target group is obtained through the preset language model.

[0044] Specifically, not all individual attribute values ​​under the preset individual attribute dimensions can be used as group attributes, that is, not all individual attribute values ​​obtained in step 102 can represent the group attributes of the target group. For example, in the target group, individual attribute values ​​with low coverage are very likely unable to represent the common attributes of the target group, or in other words, individual attribute values ​​with low coverage are usually less correlated with the group attributes of the target group. Therefore, in the above embodiment of the present application, after obtaining each individual attribute value, an attribute value screening operation is first performed, so that the preset language model mines the group attributes based on the screened attribute values ​​with good correlation with the group attributes, which can effectively improve the efficiency of attribute mining and the accuracy of the mining results.

[0045] In the embodiment of the present application, when filtering attribute values ​​from individual attribute values, the specific filtering method adopted is not limited and can be customized according to the actual situation. For example, filtering can be performed according to the coverage rate of each individual attribute value in the target group, and the attribute values ​​with a coverage rate higher than a preset coverage rate threshold are retained as the filtered attribute values; it is also possible to filter according to the semantic matching degree between each individual attribute value and the set keyword according to the set mining task keywords, and retain the attribute values ​​with a semantic matching degree higher than the preset matching degree threshold as the filtered attribute values, and so on.

[0046] Optionally, in some of the embodiments, before filtering the attribute values ​​from the individual attribute values ​​to obtain the filtered attribute values, the method further comprises:

[0047] Obtaining the attribute values ​​of the individuals in the control group under the individual attribute dimension as the control attribute values;

[0048] Calculate the information value of each individual attribute value and / or the target group index of each individual attribute value based on the individual attribute value and the control attribute value;

[0049] Filter the attribute values ​​from each individual attribute value to obtain the filtered attribute values, including:

[0050] According to the information value and / or the target group index, the attribute values ​​are screened from the individual attribute values ​​to obtain the screened attribute values.

[0051] Specifically, as mentioned above, the target group can be a group composed of multiple individuals who meet the preset mining conditions. Correspondingly, the control group in the embodiment of the present application can be a group composed of multiple individuals who do not meet the above preset mining conditions. For example, for the mining task mentioned in step 102 above: for an application that provides a certain service to users, find out what kind of people are mainly using the application, and what attributes the people who use the application usually have. If the preset mining condition is that the application has been installed or used, and the group to be mined is a user group composed of users who have installed or used the application, then the control group can be a user group composed of users who have not installed or used the application.

[0052] The information value of each individual attribute value, that is, the IV value (Information Value) of each individual attribute value. The IV value is an indicator for measuring the importance of features in statistics, which is used to identify features that have an important impact on the predicted variable. The IV value reflects the ability of the feature to distinguish different response categories. The larger the value, the more important the role of the feature in the prediction. In the embodiment of the present application, the attribute value can be screened from each individual attribute value according to the IV value of each individual attribute value to obtain the screened attribute value. The specific screening principle is not limited here. For example, an IV threshold can be preset. After calculating the IV value of the information value of each individual attribute value, the individual attribute value can be screened by the size relationship between the calculated IV value and the preset IV threshold. Exemplarily, three preset IV thresholds can be set: 0.02, 0.1 and 0.3. Individual attribute values ​​with IV values ​​greater than 0.02 are weak value attributes, individual attribute values ​​greater than 0.1 are medium value attributes, and those greater than 0.3 are strong value attributes; thereafter, the strong value attribute obtained in step 102 can be used as the screened attribute value, or both the medium value attribute and the strong value attribute can be used as the screened attribute value, and so on.

[0053] The target group index of each individual attribute value, that is, the TGI (Target Group Index) of each individual attribute value, can be used to measure the degree of inclination of a specific group relative to the overall population in a certain attribute or behavior, which can help analysts understand the attribute preferences of the target group. In the embodiment of the present application, attribute values ​​can be screened from each individual attribute value according to the TGI value of each individual attribute value to obtain the screened attribute value, and the specific screening principles are not limited here. For example: similar to the IV value, a TGI threshold value can also be preset. After calculating the TGI value of the information value of each individual attribute value, the individual attribute values ​​can be screened according to the size relationship between the calculated TGI value and the preset TGI threshold value.

[0054] In addition, in some other embodiments of the present application, when performing attribute value screening, the IV value and TGI value of each individual attribute value may also be comprehensively considered, and the specific screening principles are not limited here.

[0055] In the above embodiment of the present application, when the attribute value is screened, it is based on the perspective of information value and / or target group index. That is to say, in the above embodiment of the present application, when the attribute value is screened, the influence of each individual attribute value on the group attribute prediction is fully considered, and the attribute values ​​with less influence are deleted, while the attribute values ​​with greater influence are retained. After that, group attribute mining is performed based on the screened attribute values, which can effectively improve the efficiency of attribute mining and the accuracy of mining results.

[0056] Optionally, in some of the embodiments, performing attribute value screening from each individual attribute value to obtain screened attribute values ​​includes:

[0057] Get the filtering keywords;

[0058] Filter the attribute values ​​from each individual attribute value to obtain the filtered attribute values ​​that match the semantics of the filter keyword.

[0059] Specifically, in some embodiments of the present application, the attribute value screening can also be performed from other perspectives. For example, as described in the above embodiments, the screening keywords related to the mining task can be set, and then the filtered attribute values ​​can be obtained from the perspective of the degree of semantic matching between each individual attribute value and the above screening keywords. Since the filtered attribute values ​​can better match the screening keywords, that is, the filtered attribute values ​​can better match the specific mining tasks, therefore, group attribute mining based on the filtered attribute values ​​can also effectively improve the efficiency of attribute mining and the accuracy of mining results.

[0060] Optionally, in some of the embodiments, after obtaining the group attributes for the target group, the method further includes:

[0061] In the combined group consisting of the target group and the control group, a decision tree model is constructed based on group attributes;

[0062] According to the node value of each node in the decision tree model, the target group attributes are determined from the group attributes, and the individual capture rules are obtained by combining the target group attributes;

[0063] According to the individual capture rules, the target individuals are determined so as to promote services to the target individuals.

[0064] Specifically, in a decision tree model built based on group attributes, each node at each level corresponds to a group attribute, and the node value of the node can represent the importance of the group attribute corresponding to the node at this level in the decision-making process. Therefore, according to the node values ​​of the nodes at each level in the decision tree model, the group attributes that are more important to the decision can be determined from the group attributes as the target group attributes, and the above target group attributes can be combined to obtain the capture rules for target individual capture (commonly known as "circling people"); then, the target individual can be quickly determined according to the above rules to promote services for the target individual. The above process combines the highly interpretable group attributes with the decision tree model, which can quickly determine the individual capture rules that are easy to understand, and then promote services based on the rules, thereby effectively improving the efficiency of service promotion.

[0065] See also Figure 3 , Figure 3 Schematic diagram of constructing a decision tree model based on the group attributes obtained in the embodiment of the present application. Figure 3 Explain the construction and application process of the decision tree model:

[0066] Assuming that through the information processing solution provided in the embodiment of the present application, four group attributes are obtained: group attribute 1, group attribute 2, group attribute 3, and group attribute 4, then the following can be constructed based on the above four group attributes: Figure 3 The decision tree model shown in the figure, in which the node value of each level of the node is the IV value of the group attribute corresponding to the node at that level; Figure 3 The node values ​​of each node in the decision tree model shown can determine the target group attributes from the group attributes. Specifically, according to Figure 3 The decision tree model shown can determine that the final target group attributes are: group attribute 1 corresponding to the first-level nodes, group attribute 2 corresponding to the second-level nodes, and group attribute 3 corresponding to the third-level nodes. Then, group attribute 1 and group attribute 2 can be combined to form an individual capture rule: an individual who satisfies both group attribute 1 and group attribute 2 is the target individual; then, the target individual for service promotion can be determined from a large number of individuals according to the above rules.

[0067] Optionally, in some of the embodiments, after obtaining individual attribute values ​​of individuals included in the target group under a preset individual attribute dimension, the method further includes:

[0068] According to the preset grouping rules, each individual attribute value is divided into multiple attribute value sets;

[0069] Input each individual attribute value into the preset language model, and obtain group attribute information for the target group through the preset language model, including:

[0070] For each attribute value set, each individual attribute value in the attribute value set is input into a preset language model, and group attribute mining is performed through the preset language model to obtain group attribute information corresponding to the attribute value set;

[0071] The group attribute information corresponding to each attribute value set is combined to obtain the group attribute information for the target group.

[0072] Specifically, in the embodiment of the present application, the specific content of the preset grouping rules is not limited, and they can be set according to the attribute categories corresponding to the individual attribute dimensions. For example, age and gender can be divided into a special attribute set of individual attribute information.

[0073] In the embodiment of the present application, the individual attribute values ​​are grouped and divided in advance to form multiple attribute value sets; then, group attribute mining is performed from each attribute value set. Since the individual attribute values ​​are grouped in advance, that is, the classification process originally performed by the model is split and extracted, the overall efficiency of attribute mining can be improved.

[0074] Optionally, in some of the embodiments, after obtaining group attribute information for the target group, the method may further include:

[0075] The identification information of group attributes and the semantic description information of group attributes are input into the semantic generalization model, and the semantic description information of the target group is output through the semantic generalization model.

[0076] Specifically, in the attribute mining scenario, operators often want to have an overall understanding of the target group. Therefore, in the above embodiment of the present application, the semantic summary model is used to summarize the group attributes mined by the preset language model, so that people can have a clearer understanding of the target group. In turn, it helps to improve the user experience.

[0077] For example, you can input the following prompt words into the semantic summary model: [{"attribute classification":"xxx","attribute description":"xxx"}, ....]. The above is the group attribute of a target group. Please use no less than xx words to summarize the group. And return it in JSON format: {\"group summary\":\"xxx\"}.

[0078] For example, for the above-mentioned prompt words, the semantic summarization model can return the following result: "This group of people prefers xxx, is active in the xxx field, and has a strong purchasing tendency for xxx products. At the same time, this group of people shows rich and diverse interests in reading content, including but not limited to xxx".

[0079] See also Figure 4 , Figure 4Schematic diagram of information processing flow according to an embodiment of the present application. Figure 4 Explain the information processing process:

[0080] Step 1: construct positive and negative samples. Specifically, the target group and the control group obtain the preset individual attribute dimension information, and obtain the individual attribute values ​​of each individual in the target group and the control group under the above attribute dimension.

[0081] Step 2: Filter attribute values ​​to obtain filtered attribute values. One way is to calculate IV values ​​for the individual attribute values ​​obtained in the first step, and then filter attribute values ​​based on IV values; another way is to use a large language model to recall and mine individual attribute values ​​with the same or similar meanings as the filtered attribute values.

[0082] Step 3: Model attribute mining. Specifically, 1. Attribute value preprocessing. After obtaining the filtered attribute values, preprocess the filtered attribute values ​​first. The preprocessing method can mainly be to convert the attribute values ​​into words with clearer semantics. For example, convert the field name gender to gender, and the field value F to female. In this way, the subsequent mining results will be more stable; 2. Model attribute mining. Group the preprocessed attribute values ​​and input them into the preset language model for attribute mining to obtain group attribute information.

[0083] The fourth step is further mining and analysis. In the combined group consisting of the target group and the control group, a machine learning model (such as a decision tree model) is constructed based on the group attributes; based on the constructed machine learning model, data prediction and analysis are further performed to obtain interpretable prediction and analysis results.

[0084] Optionally, a fifth step may be performed: crowd summarization. Specifically, the identification information of the crowd attribute and the semantic description information of the crowd attribute are input into the semantic summarization model, and the semantic description information of the target crowd is output through the semantic summarization model.

[0085] Embodiment 2

[0086] See also Figure 5 , Figure 5 The information processing device provided in the embodiment of the present application includes:

[0087] The acquisition module 502 is used to acquire individual attribute values ​​of individuals in the target group under a preset individual attribute dimension;

[0088] The attribute mining module 504 is used to input each individual attribute value into a preset language model, and obtain group attribute information for the target group through the preset language model; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute.

[0089] Optionally, in some of the embodiments, the information processing device further includes:

[0090] A screening module is used to screen the attribute values ​​of the individuals in the target group under the preset individual attribute dimensions after obtaining the individual attribute values ​​of the individuals in the target group to obtain the screened attribute values;

[0091] The attribute mining module 504 is specifically used to: input the filtered attribute values ​​into a preset language model, and obtain group attribute information for the target group through the preset language model.

[0092] Optionally, in some of the embodiments, the screening module is specifically used to: before screening the attribute values ​​from each individual attribute value to obtain the screened attribute value, obtain the attribute values ​​of the individuals included in the control group under the individual attribute dimension as the control attribute value; based on the individual attribute value and the control attribute value, calculate the information value of each individual attribute value and / or the target group index of each individual attribute value; according to the information value and / or the target group index, screen the attribute values ​​from each individual attribute value to obtain the screened attribute value.

[0093] Optionally, in some of the embodiments, the information processing device further includes:

[0094] A decision tree construction module is used to construct a decision tree model based on the group attributes in a combined group consisting of the target group and the control group after obtaining the group attributes for the target group;

[0095] A rule combination module is used to determine the target group attributes from the group attributes according to the node value of each node in the decision tree model, and combine the target group attributes to obtain the individual capture rule;

[0096] The service promotion module is used to determine the target individual according to the individual capture rule, so as to promote the service to the target individual.

[0097] Optionally, in some of the embodiments, the screening module is specifically used to: obtain screening keywords; and screen attribute values ​​from each individual attribute value to obtain screened attribute values ​​that semantically match the screening keywords.

[0098] Optionally, in some of the embodiments, the information processing device further includes:

[0099] A grouping module, used to divide each individual attribute value into multiple attribute value sets according to a preset grouping rule;

[0100] The attribute mining module 504 is specifically used to: for each attribute value set, input each individual attribute value in the attribute value set into a preset language model, perform group attribute mining through the preset language model, and obtain group attribute information corresponding to the attribute value set; combine the attribute information corresponding to each attribute set to obtain group attribute information for the target group.

[0101] Optionally, in some of the embodiments, the group attribute information further includes: identification information of the group attribute; the group attribute information is generated based on semantic information of individual attribute values ​​contained in the group attribute.

[0102] Optionally, in some of the embodiments, the information processing device further includes:

[0103] The semantic summarization module is used to input the identification information of the group attributes and the semantic description information of the group attributes into the semantic summarization model after obtaining the group attribute information for the target group, and output the semantic description information of the target group through the semantic summarization model.

[0104] The information processing device of the embodiment of the present application is used to implement any of the methods in the aforementioned method embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be described in detail here. In addition, the functional implementation of each module in the information processing device of the embodiment of the present application can refer to the description of the corresponding parts in the aforementioned method embodiments, which will not be described in detail here.

[0105] Embodiment 3

[0106] Reference Figure 6 , shows a schematic diagram of the structure of an electronic device according to the third embodiment of the present application. The specific embodiment of the present application does not limit the specific implementation of the electronic device.

[0107] like Figure 6 As shown, the electronic device may include: a processor (processor) 602 , a communication interface (Communications Interface) 604 , a memory (memory) 606 , and a communication bus 608 .

[0108] in:

[0109] The processor 602 , the communication interface 604 , and the memory 606 communicate with each other via a communication bus 608 .

[0110] The communication interface 604 is used to communicate with other electronic devices or servers.

[0111] The processor 602 is used to execute the program 610, and specifically can execute the relevant steps in the above method embodiment.

[0112] Specifically, the program 610 may include program codes, which include computer operation instructions.

[0113] The processor 602 may be a CPU, or an application specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application. The one or more processors included in the smart device may be processors of the same type, such as one or more CPUs; or processors of different types, such as one or more CPUs and one or more ASICs.

[0114] The memory 606 is used to store the program 610. The memory 606 may include a high-speed RAM memory, and may also include a non-volatile memory (non-volatile memory), such as at least one disk memory.

[0115] The program 610 may be specifically used to enable the processor 602 to execute the relevant steps in the above method embodiment.

[0116] The specific implementation of each step in program 610 can refer to the corresponding description of the corresponding step in the above method embodiment, which will not be repeated here. Those skilled in the art can clearly understand that for the convenience and simplicity of description, the specific working process of the above-described devices and modules can refer to the corresponding process description in the above method embodiment, which will not be repeated here.

[0117] The present application also provides a computer storage medium on which a computer program is stored, and when the program is executed by a processor, the method described in any of the above-mentioned multiple method embodiments is implemented. The computer storage medium includes, but is not limited to: a compact disc read-only memory (CD-ROM), a random access memory (RAM), a floppy disk, a hard disk, or a magneto-optical disk, etc.

[0118] An embodiment of the present application also provides a computer program product, including computer instructions, which instruct a computing device to execute operations corresponding to any one of the above-mentioned multiple method embodiments.

[0119] In addition, it should be noted that the user-related information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to sample data used to train the model, data used for analysis, stored data, displayed data, etc.) involved in the embodiments of the present application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0120] It should be pointed out that, according to the needs of implementation, the various components / steps described in the embodiments of the present application can be split into more components / steps, or two or more components / steps or partial operations of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of the present application.

[0121] The above-mentioned method according to the embodiment of the present application can be implemented in hardware, firmware, or implemented as software or computer code that can be stored in a recording medium (such as a CD ROM, RAM, floppy disk, hard disk or magneto-optical disk), or implemented as a computer code originally stored in a remote recording medium or a non-temporary machine-readable medium downloaded through a network and to be stored in a local recording medium, so that the method described herein can be stored in such software processing on a recording medium using a general-purpose computer, a special-purpose processor or programmable or special-purpose hardware (such as an ASIC or FPGA). It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component (e.g., RAM, ROM, flash memory, etc.) that can store or receive software or computer code, and when the software or computer code is accessed and executed by a computer, a processor or hardware, the method described herein is implemented. In addition, when a general-purpose computer accesses a code for implementing the method or method shown herein, the execution of the code converts the general-purpose computer into a special-purpose computer for executing the method shown herein.

[0122] Those of ordinary skill in the art will appreciate that the units and method steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the embodiments of the present application.

[0123] The above implementation methods are only used to illustrate the embodiments of the present application, and are not limitations on the embodiments of the present application. Ordinary technicians in the relevant technical field can make various changes and modifications without departing from the spirit and scope of the embodiments of the present application. Therefore, all equivalent technical solutions also belong to the scope of the embodiments of the present application. The scope of patent protection of the embodiments of the present application should be limited by the claims.

Claims

1. An information processing method, comprising: For a target group, obtaining individual attribute values ​​of individuals in the target group under a preset individual attribute dimension; According to the preset grouping rules, each individual attribute value is divided into multiple attribute value sets; For each attribute value set, each individual attribute value in the attribute value set is input into a preset language model, and each individual attribute value is summarized and concluded by the preset language model, so that individual attribute values ​​with similar semantics are classified as the same group attribute, and a semantic interpretation of the group attribute is provided to obtain group attribute information corresponding to the attribute value set; Combining group attribute information corresponding to each attribute value set to obtain group attribute information for the target group; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute; The individual attributes represent the individual characteristics of a single individual in the target group, and the group attributes represent the common characteristics of each individual in the target group.

2. The method according to claim 1, wherein: After obtaining the individual attribute values ​​of the individuals included in the target group under the preset individual attribute dimensions, the method further includes: Filter attribute values ​​from each individual attribute value to obtain filtered attribute values; The step of inputting each individual attribute value into a preset language model and obtaining group attribute information for the target group through the preset language model includes: The filtered attribute values ​​are input into a preset language model, and group attribute information for the target group is obtained through the preset language model.

3. The method according to claim 2, wherein: Before filtering the attribute values ​​from the individual attribute values ​​to obtain the filtered attribute values, the method further includes: Obtaining attribute values ​​of individuals included in the control group under the individual attribute dimension as control attribute values; Calculate the information value of each individual attribute value and / or the target group index of each individual attribute value based on the individual attribute value and the control attribute value; The step of screening attribute values ​​from each individual attribute value to obtain screened attribute values ​​includes: According to the information value and / or the target group index, attribute values ​​are screened from the individual attribute values ​​to obtain screened attribute values.

4. The method according to claim 3, wherein: After obtaining the group attributes for the target group, the method further includes: In a combined population consisting of the target population and the control population, constructing a decision tree model based on the population attributes; According to the node value of each node in the decision tree model, determining the target group attribute from the group attributes, and combining the target group attributes to obtain the individual capture rule; According to the individual capture rule, a target individual is determined to promote services to the target individual.

5. The method according to claim 2, wherein: The step of screening attribute values ​​from each individual attribute value to obtain screened attribute values ​​includes: Get the filtering keywords; Attribute values ​​are screened from each individual attribute value to obtain screened attribute values ​​that semantically match the screening keyword.

6. The method according to any one of claims 1 to 5, wherein: The group attribute information further includes: identification information of the group attribute; the group attribute information is generated based on semantic information of individual attribute values ​​contained in the group attribute.

7. The method according to claim 6, wherein: After obtaining the group attribute information for the target group, the method further includes: The identification information of the group attributes and the semantic description information of the group attributes are input into a semantic generalization model, and the semantic description information of the target group is output through the semantic generalization model.

8. An information processing device, comprising: An acquisition module, used for acquiring individual attribute values ​​of individuals in a target group under a preset individual attribute dimension; A grouping module, used to divide each individual attribute value into multiple attribute value sets according to a preset grouping rule; The attribute mining module is used to input each individual attribute value in each attribute value set into a preset language model, summarize and conclude each individual attribute value through the preset language model, thereby classifying individual attribute values ​​with similar semantics into the same group attribute, and providing a semantic interpretation of the group attribute to obtain the group attribute information corresponding to the attribute value set; Combining group attribute information corresponding to each attribute value set to obtain group attribute information for the target group; the group attribute information includes: individual attribute values ​​contained in the group attribute, and semantic description information of the group attribute; The individual attributes represent the individual characteristics of a single individual in the target group, and the group attributes represent the common characteristics of each individual in the target group.

9. An electronic device, comprising: A processor, a memory, a communication interface and a communication bus, wherein the processor, the memory and the communication interface communicate with each other via the communication bus; The memory is used to store at least one executable instruction, and the executable instruction enables the processor to perform an operation corresponding to the method according to any one of claims 1 to 7.