Target population calculation method and apparatus, electronic device, and medium

Abstract

The application relates to the technical field of big data application, in particular to a target population calculation method and device, electronic equipment and medium, wherein the method comprises the following steps: acquiring a user identifier and an unstructured user feature value; matching the unstructured user content according to the user identifier, calculating a matching relationship between the unstructured user feature value and the unstructured user content; establishing an association relationship between the unstructured user feature value and the user identifier according to the matching relationship, and determining a population package corresponding to the unstructured user feature value based on the association relationship. Therefore, the problems that the population calculation method in the related art cannot comprehensively match the user population, leads to poor applicability, cannot meet the business requirements and the like are solved.

Description

Technical Field

[0001] This application relates to the field of big data application technology, and in particular to a method, device, electronic device and medium for calculating a target population. Background Technology

[0002] During the business process, users will generate a massive amount of user content in the business system, which will contain a lot of information about issues that business personnel care about. Business personnel need to find relevant user groups based on the keywords of the issues, conduct interviews, surveys and insight analysis, and then use data tags to perform rule-based matching with user groups to calculate user groups.

[0003] Currently, audience targeting involves combining data tags such as user attributes, order behavior, and browsing / search behavior, performing intersection, union, and difference calculations, and using code to transform visual filtering conditions into SQL (Structured Query Language) statements. This allows for the extraction of a precise user group that aligns with marketing objectives from a vast database. Audience targeting is primarily used in marketing-focused scenarios such as on-site and off-site advertising, marketing message pushes, event marketing audience verification, and membership purchases.

[0004] In related technologies, audience calculation methods often use structured user feature tags to calculate audience segments. User feature tags are structured data structures associated with a user's unique identification ID. While this approach ensures the accuracy of matching results, when calculating user audiences based on user content, it loses some user content and related audiences that are highly relevant but not completely matched with keywords. This results in a smaller matched user audience, which is difficult to meet business needs in certain situations and has poor applicability. Summary of the Invention

[0005] This application provides a target audience calculation method, device, electronic device, and medium to solve the problems in related technologies where audience calculation methods do not fully match the user audience, resulting in poor applicability and inability to meet business needs.

[0006] The first aspect of this application provides a method for calculating a target audience, comprising the following steps: obtaining a user identifier and unstructured user feature values; matching unstructured user content with the user identifier, and calculating a matching relationship between the unstructured user feature values ​​and the unstructured user content; establishing an association relationship between the unstructured user feature values ​​and the user identifier based on the matching relationship, and determining an audience package corresponding to the unstructured user feature values ​​based on the association relationship.

[0007] Optionally, calculating the matching relationship between the unstructured user feature values ​​and the unstructured user content includes: identifying preset keywords in the unstructured user feature values; vectorizing the preset keywords and user content to obtain vectorized preset keywords and user content; and calculating the matching relationship between the unstructured user feature values ​​and the unstructured user content based on the vectorized preset keywords and user content.

[0008] Optionally, before vectorizing the preset keywords and user content, the method further includes: performing word segmentation on the user content to obtain a word set after word segmentation; removing stop words from the word set after word segmentation to obtain a word set after removing stop words; and traversing the word set after removing stop words to remove duplicate words to obtain a word set after deduplication.

[0009] Optionally, the vectorized user content includes vectorized words, and the step of calculating the matching relationship between the unstructured user feature value and the unstructured user content based on the vectorized preset keywords and user content includes: calculating the similarity between the vectorized preset keywords and the vectorized words; if the similarity is greater than a matching threshold, it is determined that the unstructured user feature value and the unstructured user content have a matching relationship, otherwise it is determined that there is no matching relationship.

[0010] Optionally, establishing the association between the unstructured user feature values ​​and the user identifiers based on the matching relationship includes: if the target population is a population within the system, determining a set of target user identifiers based on the matching relationship and establishing a mapping relationship between the set of user identifiers and the user feature values; if the target population is a population uploading from outside the system, traversing and matching the uploaded set of external user identifiers and the unique identifiers of users within the system to obtain a set of target user identifiers and establishing a mapping relationship between the set of target user identifiers and the user feature values.

[0011] Optionally, after determining the audience package corresponding to the unstructured user feature value based on the association relationship, the method further includes: using the audience package to analyze the hot word distribution of user content, and performing business operations based on the hot word distribution.

[0012] Optionally, analyzing the hot word distribution of user content using the audience segment includes: calculating the user content associated with each user identifier through the user identifiers in the audience segment; performing word segmentation, stop word processing, and deduplication processing on the user content to obtain a deduplicated word set; and using the deduplicated word set to determine the hot word distribution.

[0013] A second aspect of this application provides a target audience calculation device, comprising: an acquisition module for acquiring user identifiers and unstructured user feature values; a calculation module for matching unstructured user content with the user identifiers and calculating a matching relationship between the unstructured user feature values ​​and the unstructured user content; and an establishment module for establishing an association relationship between the unstructured user feature values ​​and the user identifiers based on the matching relationship, and determining an audience package corresponding to the unstructured user feature values ​​based on the association relationship.

[0014] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the target population calculation method as described in the above embodiments.

[0015] A fourth aspect of this application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the target population calculation method as described in the above embodiments.

[0016] Therefore, this application has at least the following beneficial effects:

[0017] This application improves the matching method between unstructured user feature values ​​and user content. It mainly involves inputting unstructured user feature values ​​and matching them with unstructured user content associated with the user's unique identifier. After calculating the matching relationship, it establishes the association between the input unstructured feature values ​​and the user's unique identifier, and finally calculates the target user group associated with the input unstructured feature values. This ensures that the accuracy and applicability of the matched target group are increased, meets the business needs of business personnel, and improves work efficiency.

[0018] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0019] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein:

[0020] Figure 1 This is a flowchart of a target population calculation method provided according to an embodiment of this application;

[0021] Figure 2 This is a flowchart of a content-based method for calculating target audiences according to an embodiment of this application;

[0022] Figure 3This is a block diagram of a target population computing device according to an embodiment of this application;

[0023] Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0024] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0025] During business operations, users generate a large amount of user content in the business system, which includes a lot of information about issues that business personnel care about. Business personnel need to find relevant user groups based on the keywords of the issues to conduct interviews, surveys and insight analysis. However, the currently available user group calculation methods tend to use structured feature values ​​to perform rule-based matching with user groups, which is not suitable for calculating user groups based on unstructured user content.

[0026] In view of the above business background, the method of this application embodiment can effectively solve the technical problems in the following scenarios:

[0027] (1) Find relevant user groups based on existing keyword content; generate user groups corresponding to the same or similar content as the keyword by inputting keyword content.

[0028] (2) Analyze the distribution of high-frequency hot words based on the population within the system; conduct high-frequency keyword analysis on the population within a specified range within the system and the content they generate, generate business-related hot word distribution results, and analyze the issues and concerns of the population.

[0029] (3) Analyze the distribution of high-frequency hot words based on the uploaded audience; upload audience data containing unique user IDs from external sources and match it with all user content in the system. After matching the user range, perform high-frequency keyword analysis on the user groups in the matching results to generate business-related hot word distribution results and analyze the issues and concerns of the audience.

[0030] The target population calculation method, apparatus, electronic device, and medium of this application are described below with reference to the accompanying drawings. Specifically, Figure 1 This is a flowchart illustrating a target population calculation method provided in an embodiment of this application.

[0031] like Figure 1 As shown, the method for calculating the target population includes the following steps:

[0032] In step S101, the user identifier and unstructured user feature values ​​are obtained.

[0033] The user identifier can be a user identity ID; unstructured user feature values ​​can be relatively non-fixed user feature values, such as: not limited to user attributes, order behavior, browsing and search behavior, etc., without specific limitations.

[0034] It is understood that the embodiments of this application obtain user identifiers and unstructured user feature values ​​in order to facilitate subsequent matching of unstructured user content.

[0035] In step S102, unstructured user content is matched based on the user identifier, and the matching relationship between unstructured user feature values ​​and unstructured user content is calculated.

[0036] User content can be any text content generated by the user in the system, including but not limited to message text, work order text, audio title text, video title text, file title text, link text, etc., without specific limitations.

[0037] It is understood that the embodiments of this application match unstructured user content based on user identifiers and calculate the matching relationship between unstructured user feature values ​​and unstructured user content, so as to facilitate the subsequent determination of the audience package corresponding to the unstructured user feature values.

[0038] In this embodiment of the application, calculating the matching relationship between unstructured user feature values ​​and unstructured user content includes: identifying preset keywords in unstructured user feature values; vectorizing the preset keywords and user content to obtain vectorized preset keywords and user content; and calculating the matching relationship between unstructured user feature values ​​and unstructured user content based on the vectorized preset keywords and user content.

[0039] The preset keywords can be user-defined keywords, such as "vehicles" for browsing and searching. They can be set according to the user's actual needs, and there are no specific limitations here.

[0040] It is understood that the embodiments of this application identify keywords in unstructured user feature values, vectorize the keywords and user content, and calculate the matching relationship between unstructured user feature values ​​and user content based on the vectorized keywords and user content, so as to determine whether they belong to the target group through the matching relationship, making the matching results more accurate and meeting the work needs of business personnel.

[0041] In this embodiment of the application, before vectorizing the preset keywords and user content, the method further includes: performing word segmentation on the user content to obtain a word set after word segmentation; removing stop words from the word set after word segmentation to obtain a word set after removing stop words; and traversing the word set after removing stop words to remove duplicate words to obtain a word set after deduplication.

[0042] Among them, word segmentation is the process of breaking down user content, including text and symbols, into words that the system software can distinguish in a structured way; no specific limitations are made here.

[0043] Stop words refer to certain characters or words that are automatically filtered out in natural language data (or text) for a specific purpose, and are not specifically defined here.

[0044] It is understood that the embodiments of this application perform word segmentation on user content to obtain a word set after word segmentation. Stop words are removed from the word set after word segmentation to obtain a new word set. Duplicate words are removed from the word set after removing stop words to obtain a deduplicated word set. This allows user content to be classified so that the relevant type of user content can be quickly located when keywords are entered, thereby identifying the corresponding target audience and improving work efficiency.

[0045] In this embodiment of the application, the vectorized user content includes vectorized words. The matching relationship between unstructured user feature values ​​and unstructured user content is calculated based on the vectorized preset keywords and user content, including: calculating the similarity between the vectorized preset keywords and vectorized words; if the similarity is greater than the matching threshold, it is determined that there is a matching relationship between the unstructured user feature values ​​and unstructured user content, otherwise it is determined that there is no matching relationship.

[0046] Similarity refers to the vector distance between the vectors corresponding to two words. Each word is uniquely associated with a vector [v1, v2], for example, word A corresponds to vector [a1, a2] and word B corresponds to vector [b1, b2]. The vector distance between the two words is the similarity. The higher the matching degree, the more matched the two words are considered, S∈[0,1].

[0047] The matching threshold can be a threshold set by the user in advance, and no specific restrictions are made here.

[0048] It is understood that the embodiments of this application calculate and compare the similarity between vectorized keywords and words. When the similarity is greater than the matching threshold, it is determined that there is a matching relationship between the unstructured user feature value and the user content. Otherwise, there is no matching relationship, so that users can find the target group with a high degree of matching when they input keywords.

[0049] In step S103, an association relationship between unstructured user feature values ​​and user identifiers is established based on the matching relationship, and the population package corresponding to the unstructured user feature values ​​is determined based on the association relationship.

[0050] It is understood that the embodiments of this application establish an association between unstructured user feature values ​​and user identifiers through matching relationships, and then determine the population package corresponding to the unstructured user feature values ​​based on the association relationship. This allows for the rapid location of the target user population, ensuring the accuracy of the search, expanding the scope of application, meeting the work needs of business personnel, and improving work efficiency.

[0051] In this embodiment of the application, establishing an association between unstructured user feature values ​​and user identifiers based on matching relationships includes: if the target population is a population within the system, determining a set of target user identifiers based on the matching relationship and establishing a mapping relationship between the set of user identifiers and user feature values; if the target population is a population uploading from outside the system, traversing and matching the uploaded set of external user identifiers and the unique identifiers of users within the system to obtain a set of target user identifiers and establishing a mapping relationship between the set of target user identifiers and user feature values.

[0052] The traversal matching can be performed by matching each user identifier in the external user identifier set with the unique identifier in the system one by one, without any specific limitations.

[0053] It is understood that, in this embodiment of the application, if the target audience is a group within the system, the user identifier set is determined based on the matching relationship, and a mapping relationship is established between the user identifier combination and the user feature value; if the target audience is an external user group, the uploaded external user identifier combination and the internal user identifier are traversed and matched to obtain the target user set, and a mapping relationship is established between the target user identifier set and the user feature value, so that users can quickly locate the target audience type when searching, thereby improving work efficiency.

[0054] In this embodiment of the application, after determining the audience package corresponding to the unstructured user feature value based on the association relationship, the method further includes: using the audience package to analyze the hot word distribution of user content, and performing business operations based on the hot word distribution.

[0055] Hot words refer to the most frequent words in user-generated content, sorted by frequency of occurrence, without any specific definition.

[0056] It is understood that the embodiments of this application utilize audience segmentation to analyze the distribution of hot words in user content and perform business operations based on the distribution of hot words, so that business personnel can analyze the concerns of the audience based on the distribution of hot words in user content calculated from the input audience segment, quickly find the corresponding needs of the audience, and improve work efficiency.

[0057] In this embodiment of the application, the hot word distribution of user content is analyzed using audience segments, including: calculating the user content associated with each user identifier through the user identifiers in the audience segment; performing word segmentation, stop word processing, and deduplication processing on the user content to obtain a deduplicated word set; and using the deduplicated word set to determine the hot word distribution.

[0058] It is understood that, in this embodiment of the application, user content associated with each user identifier is calculated by using user identifiers within the audience package. The user content is then processed by word segmentation, stop word processing, and deduplication to obtain a deduplicated word set. The distribution of hot words is determined using the deduplicated word set, which facilitates the subsequent analysis of the target audience's concerns. This allows for the rapid location of the target audience during subsequent keyword searches, thereby improving work efficiency.

[0059] The target user group calculation method proposed in this application obtains user identifiers and unstructured user feature values, matches unstructured user content with user identifiers, calculates the matching relationship between unstructured user feature values ​​and unstructured user content, establishes an association between unstructured user feature values ​​and user identifiers through the matching relationship, and then determines the user group corresponding to the unstructured user feature values ​​based on the association relationship. This allows for rapid location of the target user group, ensuring accuracy, expanding the scope of application, meeting the work needs of business personnel, and improving work efficiency. Therefore, it solves the problems of incomplete matching of user groups in related technologies, resulting in poor applicability and inability to meet business needs.

[0060] The following will combine Figure 2 The method for calculating the target population is explained in detail, and the specific process is as follows:

[0061] First, the user inputs feature keywords, and the system checks whether the user content has been processed. If it has been processed, the word vectorization module is used to vectorize the input keywords and user content.

[0062] Then, the vectorized keywords and user content are output. The similarity calculation module calculates the matching relationship between the input keywords and user content and outputs the user content words with matching relationship. The audience package calculation module calculates the set of user IDs associated with the user content and outputs the audience package IDs associated with the user content words.

[0063] Finally, the audience segmentation application is applied: If the user content in the system has not been processed, the word set is input into the stop word processing module, which outputs the word set after removing stop words. The output word set is then traversed and deduplicated, and the deduplicated word set is output. Then, the word vectorization module is used to vectorize the input keywords and user content, and the vectorized keywords and user content are output. The similarity calculation module calculates the matching relationship between the input keywords and user content and outputs the user content words with matching relationships. The audience segmentation calculation module calculates the set of user IDs associated with the user content and outputs the audience segment IDs associated with the user content words. Finally, the audience segmentation application is applied.

[0064] The target population calculation method will be further illustrated through the following three embodiments, which will be described from the following four aspects:

[0065] 1. User content processing: By segmenting user content text within the system into words and then vectorizing it, a word set is generated;

[0066] 2. Matching Relationship Calculation: The input feature value keywords are used to calculate the matching relationship between the input feature value keywords and user content in the system through vectorized calculation.

[0067] 3. Audience Package Calculation: After inputting feature keywords, a set of unique identification IDs of all users whose content is associated with the input feature keywords is established as the user group associated with the input feature keywords, and a unique audience package ID is assigned to the user group.

[0068] 4. Audience Segmentation Application: Business personnel can use the audience segment ID calculated from the input keywords to perform business operations on the unique user identification ID corresponding to the audience segment ID, such as calculating the distribution of hot keywords in user content, marketing placement, interview surveys, and user profile insights.

[0069] Based on the above four aspects, the specific implementation plan is as follows:

[0070] Example 1: Calculating user groups based on user content

[0071] 1. User content processing

[0072] 1) Input the user content in the system as the text to be processed into the word segmentation module;

[0073] 2) The word segmentation module processes user content and outputs a set of segmented words;

[0074] 3) Input the word set into the stop word processing module, and output the word set after removing stop words;

[0075] 4) Iterate through the word set output in step 3) to remove duplicates, and output the deduplicated word set.

[0076] The word segmentation module can decompose continuous text content into a set of words that the system can distinguish. The word segmentation module uses the commonly used jieba word segmentation algorithm. The jieba algorithm uses a prefix dictionary to achieve efficient word graph scanning, generating a directed acyclic graph (DAG) composed of all possible word combinations generated by Chinese characters in a sentence. Then, dynamic programming is used to find the maximum probability path and identify the maximum segmentation combination based on word frequency. For out-of-vocabulary words, an Hidden Markov Model (HMM) based on the word-forming ability of Chinese characters is used, employing the Viterbi algorithm.

[0077] The stop word module automatically filters certain characters or words from natural language data (or text) for a specific purpose. It utilizes the commonly used Baidu stop word dictionary. By performing regular expression matching on the input word set and the words in the stop word dictionary, words that match the results are removed from the word set, and the remaining word set is output as the final result.

[0078] 2. Matching Relationship Calculation

[0079] 1) Input the keywords whose matching relationship needs to be calculated and the words in the word set corresponding to the user content into the word vectorization module, and output the vector value of each word;

[0080] 2) Input the vectors corresponding to the keywords whose matching relationship is to be calculated and the vectors corresponding to the user content words into the similarity calculation module, and output the similarity results;

[0081] 3) Compare the similarity calculation result with the similarity threshold, and output the matching relationship result between the two words. The matching relationship includes: yes or no. Two keywords with a similarity greater than or equal to the similarity threshold are considered to have a matching relationship.

[0082] The matching degree calculation module is implemented using the commonly used word2vec algorithm and the Chinese Wikipedia corpus.

[0083] In this module, words from Chinese Wikipedia are used as training corpora and input into word2vec to train word vectors. The word2vec algorithm, also known as word embeddings, converts words in natural language into dense vectors that computers can understand. Word2vec mainly has two modes: CBOW (Continuous Bag of Words) and Skip-Gram. CBOW infers target words from the original sentence; Skip-Gram, on the other hand, infers the original sentence from the target words. CBOW is more suitable for small databases, while Skip-Gram performs better on large corpora. Here, we choose the CBOW model to train the corpus, specifically the Chinese Wikipedia corpus.

[0084] In the similarity calculation module, similarity is defined as the vector distance between the vectors corresponding to two words. Each word is uniquely assigned a vector [v(1), v(2)]. For example, the vector corresponding to word A is [a(1), a(2)], and the vector corresponding to word B is [b(1), b(2)]. The vector distance between the two words, i.e., the similarity, is S∈[0,1]. The higher the similarity, the closer the meanings of the two words are considered to be.

[0085] 3. Crowd Package Calculation

[0086] 1) After calculating the user content that matches the input keywords, input all matching content into the audience segment calculation module;

[0087] 2) The audience segmentation module calculates a unique set of IDs for all users associated with the user content. This set represents the user audience that matches the input keywords.

[0088] 3) Mark the group segment with a unique identifier and map it to a unique group segment identification ID.

[0089] 4) Establish a mapping relationship between the input keywords and the calculated audience package IDs.

[0090] The audience segmentation module can calculate all unique user IDs associated with a specific user characteristic and its value by inputting that characteristic. For example, it can calculate the unique user ID for the user who generated the content, or it can calculate the unique user ID for the user's age. User content and age are examples of user characteristics. After calculating all unique user IDs associated with the characteristic value, the module removes duplicates from all unique user IDs and outputs them as a set.

[0091] 4. Crowd Package Application

[0092] Subsequently, business personnel can use the audience segment ID calculated from the input keywords to perform business operations on the unique user identification ID corresponding to the audience segment ID, such as marketing campaigns, interviews and surveys, and user profile insights.

[0093] Example 2: Hot word distribution of user content based on in-system demographic analysis

[0094] 1. Crowd Package Calculation

[0095] The system uses a crowd packet calculation module to calculate the set of crowd packets that are associated with the input feature values.

[0096] 2. User content calculation

[0097] 1) Calculate the user content associated with each ID by using the unique user identification ID within the crowd package.

[0098] 2) Input the user content associated with all IDs calculated in step 1) as the text to be processed into the word segmentation module;

[0099] 3) The word segmentation module processes user content and outputs a set of segmented words;

[0100] 4) Input the word set into the stop word processing module and output the word set after removing stop words;

[0101] 5) Iterate through the word set output in step 4) to remove duplicates, and output the deduplicated word set.

[0102] 3. Hot word distribution calculation

[0103] The main process involves inputting the word set output in step 2 into the hot word calculation module, which then outputs a hot word distribution ranking data structure. The data structure includes: {word content}{word occurrence count}{word ranking order (sorted in descending order of occurrence count)}.

[0104] The hot word calculation module can count the number of times each word appears in the word set, and sort them in descending order based on the number of occurrences. It outputs the ranking and the number of occurrences of each word. The ranking is based on the number of occurrences increasing from 1. The more occurrences, the closer the ranking is to 1. If the number of occurrences is the same, the ranking is the same.

[0105] 4. Application of hot word distribution

[0106] The main purpose is that subsequent business personnel can use the distribution of hot keywords in user content calculated from the input audience package to carry out business operations, such as user profile insights, interviews and surveys, etc.

[0107] Scenario 3: Analyzing the distribution of hot keywords in user content based on the uploaded user group

[0108] 1. Crowd Package Calculation

[0109] 1) Upload data, including a set of unique user identification IDs, from outside the system to the system;

[0110] 2) Iterate through and match the uploaded set of external user unique identifiers with the existing user unique identifiers in the system to calculate all user unique identifiers in the system that can be accurately matched;

[0111] 3) Output the set of all unique user IDs calculated in step 2) as the matching result of the uploaded audience.

[0112] 2. User content calculation

[0113] 1) Calculate the user content associated with each ID using the unique user identification ID within the crowd package output in step 1;

[0114] 2) Input the user content associated with all IDs calculated in step 1) as the text to be processed into the word segmentation module;

[0115] 3) The word segmentation module processes user content and outputs a set of segmented words;

[0116] 4) Input the word set into the stop word processing module and output the word set after removing stop words;

[0117] 5) Iterate through the word set output in step 4) to remove duplicates, and output the deduplicated word set.

[0118] 3. Hot word distribution calculation

[0119] Input the word set output in step 2 into the hot word calculation module, and output the hot word distribution ranking data structure. The data structure includes: {word content}{word occurrence count}{word ranking order (sorted in descending order of occurrence count)}

[0120] 4. Application of hot word distribution

[0121] Business personnel can then use the distribution of hot keywords in user content calculated from the input audience segment to perform business operations, such as user profile insights and interview surveys.

[0122] In summary, by inputting unstructured user feature values, matching and calculating them with unstructured user content associated with the user's unique identifier, and establishing the association between the input unstructured feature values ​​and the user's unique identifier, the target user group associated with the input unstructured feature values ​​can be calculated. This ensures that the accuracy and applicability of the matched target group are increased, meets the business needs of business personnel, and improves work efficiency.

[0123] Next, the target population calculation device proposed according to the embodiments of this application is described with reference to the accompanying drawings.

[0124] Figure 3 This is a block diagram of a target population computing device according to an embodiment of this application.

[0125] like Figure 3 As shown, the target population calculation device 10 includes: an acquisition module 100, a calculation module 200, and an establishment module 300.

[0126] The acquisition module 100 is used to acquire user identifiers and unstructured user feature values; the calculation module 200 is used to match unstructured user content with user identifiers and calculate the matching relationship between unstructured user feature values ​​and unstructured user content; the establishment module 300 is used to establish the association relationship between unstructured user feature values ​​and user identifiers based on the matching relationship and determine the audience package corresponding to the unstructured user feature values ​​based on the association relationship.

[0127] It should be noted that the foregoing explanation of the target population calculation method embodiment also applies to the target population calculation device of this embodiment, and will not be repeated here.

[0128] The target user group calculation device proposed in this application obtains user identifiers and unstructured user feature values, matches unstructured user content with user identifiers, calculates the matching relationship between unstructured user feature values ​​and unstructured user content, establishes an association between unstructured user feature values ​​and user identifiers through the matching relationship, and then determines the user group corresponding to the unstructured user feature values ​​based on the association relationship. This allows for rapid location of the target user group, ensuring accuracy, expanding the scope of application, meeting the work needs of business personnel, and improving work efficiency. Therefore, it solves the problems of incomplete matching of user groups in related technologies, resulting in poor applicability and inability to meet business needs.

[0129] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include:

[0130] The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0131] When the processor 402 executes the program, it implements the target population calculation method provided in the above embodiments.

[0132] Furthermore, electronic devices also include:

[0133] Communication interface 403 is used for communication between memory 401 and processor 402.

[0134] The memory 401 is used to store computer programs that can run on the processor 402.

[0135] The memory 401 may include high-speed RAM (Random Access Memory) memory, and may also include non-volatile memory, such as at least one disk storage.

[0136] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus, etc. The bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0137] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0138] Processor 402 may be a CPU (Central Processing Unit), an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement embodiments of this application.

[0139] This application also provides a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the above-described target population calculation method.

[0140] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0141] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0142] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0143] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (FPGAs), field-programmable gate arrays (FPGAs), etc.

[0144] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0145] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for calculating a target population, characterized in that, Includes the following steps: Obtain user identifiers and unstructured user feature values; Match unstructured user content based on the user identifier, and calculate the matching relationship between the unstructured user feature value and the unstructured user content; Based on the matching relationship, establish the association between the unstructured user feature value and the user identifier, and determine the audience package corresponding to the unstructured user feature value based on the association relationship; The calculation of the matching relationship between the unstructured user feature values ​​and the unstructured user content includes: Identify preset keywords in the unstructured user feature values; The preset keywords and user content are vectorized to obtain vectorized preset keywords and user content. Based on the vectorized preset keywords and user content, the matching relationship between the unstructured user feature value and the unstructured user content is calculated. Before vectorizing the preset keywords and user content, the process also includes: The user content is segmented into words to obtain a set of segmented words; Remove the stop words from the word set after word segmentation to obtain the word set after removing stop words; Traverse the set of words after removing stop words and remove duplicate words to obtain a deduplicated set of words; The vectorized user content includes vectorized words. The step of calculating the matching relationship between the unstructured user feature values ​​and the unstructured user content based on the vectorized preset keywords and user content includes: Calculate the similarity between the vectorized preset keywords and the vectorized words; If the similarity is greater than the matching threshold, it is determined that there is a matching relationship between the unstructured user feature value and the unstructured user content; otherwise, it is determined that there is no matching relationship.

2. The method according to claim 1, characterized in that, The step of establishing the association between the unstructured user feature value and the user identifier based on the matching relationship includes: If the target audience is within the system, then the target user identifier set is determined based on the matching relationship, and a mapping relationship is established between the user identifier set and the user feature value; If the target audience is an external user group, the uploaded external user identifier set and the internal user unique identifier are traversed and matched to obtain the target user identifier set, and a mapping relationship is established between the target user identifier set and the user feature value.

3. The method according to claim 1, characterized in that, After determining the user demographics corresponding to the unstructured user feature values ​​based on the aforementioned association, the process further includes: The distribution of hot words in user content is analyzed using the aforementioned audience segments, and business operations are performed based on this distribution.

4. The method according to claim 3, characterized in that, Analyzing the distribution of hot words in user content using the aforementioned audience segmentation includes: Calculate the user content associated with each user identifier within the audience package; The user content is processed by word segmentation, stop word processing, and deduplication to obtain a deduplicated word set. The distribution of hot words is determined using the deduplicated set of words.

5. A target population calculation device, characterized in that, include: The acquisition module is used to acquire user identifiers and unstructured user feature values; The calculation module is used to match unstructured user content based on the user identifier and calculate the matching relationship between the unstructured user feature value and the unstructured user content. A module is established to establish an association between the unstructured user feature values ​​and the user identifier based on the matching relationship, and to determine the audience package corresponding to the unstructured user feature values ​​based on the association relationship. The calculation of the matching relationship between the unstructured user feature values ​​and the unstructured user content includes: Identify preset keywords in the unstructured user feature values; The preset keywords and user content are vectorized to obtain vectorized preset keywords and user content. Based on the vectorized preset keywords and user content, the matching relationship between the unstructured user feature value and the unstructured user content is calculated. Before vectorizing the preset keywords and user content, the process also includes: The user content is segmented into words to obtain a set of segmented words; Remove the stop words from the word set after word segmentation to obtain the word set after removing stop words; Traverse the set of words after removing stop words and remove duplicate words to obtain a deduplicated set of words; The vectorized user content includes vectorized words. The step of calculating the matching relationship between the unstructured user feature values ​​and the unstructured user content based on the vectorized preset keywords and user content includes: Calculate the similarity between the vectorized preset keywords and the vectorized words; If the similarity is greater than the matching threshold, it is determined that there is a matching relationship between the unstructured user feature value and the unstructured user content; otherwise, it is determined that there is no matching relationship.

6. An electronic device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the target population calculation method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the target population calculation method as described in any one of claims 1-4.

Images