Potential user identification method and related device

Abstract

The invention discloses a potential user identification method and related equipment, and relates to the technical field of data mining, and the method comprises the steps: obtaining a text crawled by using a keyword of a commodity, carrying out the semantic analysis of the text, obtaining description tags, further screening the description tags, obtaining a matched target circling condition by using the target description tags, and carrying out the screening of the target circling condition. Therefore, potential users are identified. According to the method and the device, the potential user can be identified by using the text, and the problem that the potential user cannot be accurately identified when related data of query, browsing, collection and purchase behaviors of the user are lacked in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of data mining technology, and in particular to a potential user identification method and related devices. Background Art

[0002] With the development of e-commerce, more and more people are shopping on shopping websites or apps.

[0003] In order to achieve precision marketing, it is necessary to accurately identify potential users. Currently, it is common to analyze whether a user is a potential user of a certain product or a certain category of products based on the user's query, browsing, collection, and purchase behaviors. For example: if a user has searched for a certain product, the user is determined to be a potential user of the product. Another example: if a user browses multiple different products under a certain category of products, the user is determined to be a potential user of the product.

[0004] However, in some cases, the relevant data of users' query, browsing, collection, and purchase behaviors cannot be obtained. For example, the data generated by users' query, browsing, collection, and purchase of medical products may involve the user's health information, which is relatively sensitive. The shopping service provider has kept the data generated by users' query, browsing, collection, and purchase of medical products confidential, making it impossible for third parties to obtain this data. The lack of the relevant data of users' query, browsing, collection, and purchase behaviors has resulted in the inability of existing technologies to accurately identify potential users. Summary of the invention

[0005] In view of the above problems, the present application provides a potential user identification method and related devices to achieve the purpose of accurately identifying potential users. The specific scheme is as follows:

[0006] The first aspect of the present application provides a potential user identification method, comprising:

[0007] Obtaining multiple texts crawled from the network according to target keywords, wherein the target keywords are keywords of target products;

[0008] Performing semantic analysis on part or all of the content in each of the texts, and adding description tags to each of the texts based on the semantic analysis results;

[0009] Filtering out a target description tag from the description tags according to the number of the texts with the description tags;

[0010] Filtering out a target selection condition that matches the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting a user on a shopping platform;

[0011] At least some of the users who meet the target selection conditions are determined as potential users of the target product.

[0012] In a possible implementation, the performing semantic analysis on part or all of the content in each of the texts includes:

[0013] Segmenting each of the texts to obtain a plurality of words, and screening out crowd type-related words and shopping demand-related words from the plurality of words;

[0014] Constructing a phrase including vocabulary associated with a crowd type and / or vocabulary associated with a shopping demand, wherein the phrase includes a verb;

[0015] The phrase is semantically analyzed.

[0016] In a possible implementation, the performing semantic analysis on part or all of the content in each of the texts includes:

[0017] Using a preset rule matching algorithm, identifying phrases in the text that conform to the preset rule matching algorithm;

[0018] The phrase is semantically analyzed.

[0019] In a possible implementation, after filtering out a target description tag from each of the description tags according to the number of the texts with each of the description tags, the method further includes:

[0020] Verifying the target description label using multi-channel information;

[0021] The step of selecting a target selection condition that matches the target description tag from a preset user selection condition library includes:

[0022] The target selection condition that matches the verified target description tag is screened out from the preset user selection condition library.

[0023] In a possible implementation, determining at least some of the users who meet the target selection condition as potential users of the target product includes:

[0024] Predicting at least one shopping-related indicator for each user who meets the target selection condition;

[0025] Potential users of the target product are screened out from the users according to the shopping association index of each user.

[0026] In a possible implementation, after determining at least some of the users who meet the target selection condition as potential users of the target product, the method further includes:

[0027] Send a test advertisement to at least some of the potential users to obtain a test result.

[0028] A second aspect of the present application provides a potential user identification method and device, comprising:

[0029] An obtaining unit, used to obtain a plurality of texts crawled from the network according to a target keyword, wherein the target keyword is a keyword of a target product;

[0030] An analysis unit, configured to perform semantic analysis on part or all of the content in each of the texts, and to add a description tag to each of the texts based on the semantic analysis result;

[0031] A first screening unit, configured to screen out a target description tag from each of the description tags according to the number of the texts with each of the description tags;

[0032] A second screening unit, configured to screen out a target selection condition that matches the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting a user on a shopping platform;

[0033] The determination unit is used to determine at least some of the users who meet the target selection conditions as potential users of the target product.

[0034] In a third aspect, the present application provides a computer program product, including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements the potential user identification method of the first aspect or any implementation of the first aspect. In a fourth aspect, the present application provides an electronic device, including at least one processor and a memory connected to the processor, wherein:

[0035] The memory is used to store computer programs;

[0036] The processor is used to execute the computer program so that the electronic device can implement the potential user identification method of the first aspect or any implementation manner of the first aspect.

[0037] A fifth aspect of the present application provides a computer storage medium, which carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can use the potential user identification method of the first aspect or any implementation of the first aspect.

[0038] By means of the above technical solution, the potential user identification method and related equipment provided by the present application can obtain text crawled by using keywords of commodities, and perform semantic analysis on the text to obtain description tags, further screen the description tags, and obtain matching target selection conditions by using target description tags, thereby identifying potential users. The present application can identify potential users by using text, solving the problem that the prior art cannot accurately identify potential users when the relevant data of user query, browsing, collection, and purchase behavior is missing. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. Throughout the accompanying drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and the originals and elements are not necessarily drawn to scale.

[0040] Figure 1 A flowchart of a potential user identification method provided in an embodiment of the present application;

[0041] Figure 2 A schematic diagram of the structure of a potential user identification device provided in an embodiment of the present application;

[0042] Figure 3 A schematic diagram of the hardware structure of a potential user identification device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0043] The following describes the embodiments of the present application in conjunction with the drawings in the embodiments of the present application. The terms used in the implementation method section of the present application are only used to explain the specific embodiments of the present application, and are not intended to limit the present application.

[0044] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0045] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and need not be used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, which is only to describe the distinction mode adopted by the objects of the same attributes when describing in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0046] In order to achieve the purpose of accurately identifying potential users, the present application provides a potential user identification method. The potential user identification method provided by the present application is further described in detail below in conjunction with the accompanying drawings and specific implementation methods.

[0047] Please refer to the attached Figure 1 , Figure 1 A flowchart of a potential user identification method provided in an embodiment of the present application. The method may include the following steps:

[0048] Step S101: obtaining a plurality of texts crawled from the network according to a target keyword, where the target keyword is a keyword of a target product.

[0049] It should be noted that target keywords are words used to describe or refer to a certain type of goods or services, and can specifically refer to "keywords of target goods", that is, keywords directly related to the goods you want to analyze, study or promote. Web crawling (also known as web crawlers, data crawling) is an automated program used to collect data from the Internet and save it locally or in a database. Text refers to various information carriers related to target keywords obtained through web crawling, such as product reviews, social media posts, video subtitles, etc., which can contain descriptions, evaluations, user experience and other information of the target products.

[0050] It should be noted that the text crawled in step S101 of the present application is a text that can be crawled legally, such as: a public text on the Internet.

[0051] In this application, you can search for web pages based on preset target keywords and crawl multiple texts from the web pages. For example, search for "blood glucose meter" on e-commerce platform A, crawl reviews for the top 30 blood glucose meter products, and get "For my dad, he has always used this brand of blood glucose meter" and so on.

[0052] Furthermore, specified symbols and blank characters in the text can be deleted to ensure the neatness of the text, and some sentences in the text that are suspected to be copied and pasted in batches can be identified and removed to reduce the redundancy and noise of the text.

[0053] Step S102: semantic analysis is performed on part or all of the content in each text, and description tags are added to each text based on the semantic analysis results.

[0054] In this application, the content of the text can be deeply analyzed to accurately add a description tag to each text, which helps to quickly understand and classify the text content later.

[0055] As an implementable method, each text may be segmented to obtain multiple words, and crowd type-related words and shopping demand-related words may be screened out from the multiple words; phrases containing crowd type-related words and / or shopping demand-related words may be constructed, wherein the phrases include verbs; and semantic analysis may be performed on the phrases.

[0056] It should be noted that crowd type-related words usually describe or refer to different crowd types, including words that directly indicate crowd types and words that indirectly indicate crowd types. Shopping demand-related words are usually related to shopping behavior or purchase intention, such as "buy", "discount", "discount", etc. These words can reveal users' shopping needs or consumption preferences.

[0057] Among them, the words that directly indicate the type of people may include: "elderly", "young people", "obesity", "gift giving", "high blood pressure", etc. The words that indirectly indicate the type of people may include: "weight loss", "slimming", "low fat", etc. for obese people; "dad", "mom", etc. for gift giving people; "blood pressure", "blood pressure reduction", etc. for high blood pressure people; "sweet", "chocolate", "candy", "dessert", etc. for sweets lovers; "examination" for physical examination people. These words help to identify the target audience or user groups involved in the text.

[0058] Specifically, the text is segmented using a word segmentation tool or algorithm to break the text into independent vocabulary units. For example, the text "I bought it for my dad, he has always used this brand" will be identified as "for", "dad", "of", "he", "always use", "this brand", "blood glucose meter" and other words after word segmentation.

[0059] We can further analyze the frequency of occurrence of words and filter out words that are closely related to the type of people and purchasing needs. These words usually have high semantic value and can reflect the core content of the text. For example, "dad" is a word related to the type of people, and "give" is a word related to shopping needs.

[0060] The selected words can be used in a variety of ways to construct phrases with clear semantics, which can more accurately reflect the content of the text. These phrases not only contain words that describe the type of people and shopping needs, but also add verbs, nouns or adjectives, so that the phrases can express more complete semantic information. For example, you can combine the two words "give to" in the verb library and the selected word "dad" to construct a phrase like "give to dad". Alternatively, the selected word "dad" is input into a pre-trained language model, and the output is "give to dad".

[0061] The constructed phrases can be semantically analyzed to dig deeper into the meaning and intent behind the phrases. Semantic analysis can be achieved using pre-trained language models, word embedding technology, or deep learning methods in natural language processing. These technologies can capture the semantic features in the phrases and obtain semantic analysis results.

[0062] As another possible implementation method, a preset rule matching algorithm may be used to identify phrases in the text that meet the preset rule matching algorithm; and semantic analysis may be performed on the phrases.

[0063] Specifically, a preset rule matching algorithm is used to identify phrases in the text that meet a specific pattern. The preset rule matching algorithm mentioned here can be the structure of "whether there is a verb plus a noun in the text". First, the text can be divided into separate words or phrases, and a part-of-speech tag (such as verb, noun) is assigned to each word. Then, each word and its subsequent words (or words within a specific range) are scanned to determine whether they form a "verb + noun" structure, and these phrases are semantically analyzed.

[0064] The preset rule matching algorithm can also be a structure of "whether there is a verb plus an adjective plus a noun in the text". First, the text can be divided into separate words or phrases, and a part-of-speech tag (such as verb, noun, adjective) is assigned to each word. Then, each word and its subsequent words (or words within a specific range) are scanned to determine whether they form a "verb + adjective + noun" structure, and these phrases are semantically analyzed.

[0065] Since the text contains a lot of content and is complex, it may contain some content that is irrelevant to determining the type of people or determining the purchasing needs. If the text is directly semantically analyzed, these irrelevant contents may interfere with the results of the semantic analysis. Therefore, the phrases of this application are determined based on the text and are related to determining the type of people or determining the purchasing needs. Performing semantic analysis on the phrases can avoid interference.

[0066] It should be noted that the description tag refers to a tag used to describe the user, for example, describing the user's interests, identity, shopping purposes, etc.

[0067] A tag library containing multiple descriptive tags needs to be built in advance. These tags should be able to cover the topics or categories that the text may involve. The construction of the tag library can be based on multiple sources such as domain knowledge, user feedback, historical data, etc. Convert the results of semantic analysis into vector representations for numerical calculations. This can be achieved through methods such as bag of words model, TF-IDF, word embedding (such as Word2Vec, BERT, etc.). Similarly, convert the tags into vector representations. This can be achieved by vectorizing the vocabulary in the tag and then taking the average or performing other aggregation operations. Use metrics such as cosine similarity, Euclidean distance, Jaccard similarity, etc. to calculate the similarity between the semantic analysis result vector and the tag vector. Based on the calculation results of semantic similarity, select the tag with the highest similarity as the corresponding tag of the text. This usually involves setting a threshold, and only when the similarity exceeds the threshold, the tag is assigned to the text. If the similarity of multiple tags is high, the one with the highest similarity can be selected.

[0068] Since users use different words to express their needs, the same group of people may be labeled differently. Therefore, we can measure the correlation between labels in pairs, quantify the correlation between labels, and merge highly correlated labels. For example, to measure the correlation between "send parents" and "send relatives": Among all the texts corresponding to the "send parents" label, the proportion of texts mentioning "family" to the total number of texts is The proportion of texts mentioning "gift giving" to all texts is Similarly, among all the texts corresponding to the label "send to relatives", the proportion of texts mentioning "family" to all texts is The proportion of texts mentioning "gift giving" to all texts is , calculate the Euclidean distance: , the smaller the Euclidean distance, the higher the relevance. The higher the relevance, the more likely it is that two situations will occur: more texts are tagged with both labels at the same time, and in fact the texts come from the same person, or the two groups of people discuss similar words around the product and have very similar needs. Therefore, it is meaningless to distinguish them into two labels, and it is more appropriate to merge them into "gift-giving group". At this time, the labels of all texts corresponding to the "gift-giving parents" label and all texts corresponding to the "gift-giving relatives" label are updated to "gift-giving group", which can keep the number of labels within 30 groups.

[0069] Step S103: Filter out target description tags from the description tags according to the number of texts with each description tag.

[0070] In this application, the number of times each group of tags is mentioned, that is, the number of texts tagged with the tag, can be counted, and the 15 groups of tags with the highest number of mentions can be screened out as target description tags. Each group of target description tags reflects a group of potential users who have demand for the product. For example, the group behind the "gift-giving crowd" label may be users who care about the health of their relatives. These users buy blood glucose meters because their relatives need to prevent or test blood sugar. The group behind the "sweet tooth lovers" label may be users who like to eat sweets or high-carb foods. These users need blood glucose meters to test their blood sugar.

[0071] Step S104: Filter out target selection conditions that match the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting users on the shopping platform.

[0072] In this application, in order to map the target description tag to the shopping platform, different shopping platforms provide a user selection condition library for selecting population packages. The user selection condition library includes all available user selection conditions on each shopping platform. These conditions usually include multiple dimensions such as the user's purchase history, comment history, search behavior, reading interests, and life interests. For example: E-commerce platform B uses crawler technology to obtain all selectable conditions and their definitions on platforms such as "Data Engine" and "Dharma Pan" to establish a user selection condition library. Among them, the definition of "life interests" is based on the behavior in different scenarios on Taobao Mobile, and abstracts characteristic populations related to life interests from multiple angles such as style / life stage / purchasing power / interest preferences, such as "cup control".

[0073] Natural language processing technology can be used to match a group of potential users who have demand for products behind each set of target description labels with the definition of selectable conditions. Then the semantic similarity calculation in natural language processing technology is used to assist in judging fuzzy matching. For example, "gift-giving people" may be mapped as "gift-giving people for personal care / health care / massage equipment for elders" on e-commerce platform B, and may be mapped as "gift-giving people" on e-commerce platform A. For "sweet lovers", they may be mapped as buying snacks (chocolate, cake) on e-commerce platform B, and may be mapped as "(sweet) keyword search behavior" on e-commerce platform A, etc.

[0074] If some target description tags cannot find matching selectable conditions on shopping platforms, such as "drug poisoning population", these tags may not be suitable for operation on e-commerce platforms. Such people are more suitable to be reached through marketing methods other than shopping platforms.

[0075] Step S105: determining at least some of the users who meet the target selection criteria as potential users of the target product.

[0076] In the present application, at least one shopping-related index of each user who meets the target selection criteria can be predicted; and potential users of the target product can be screened out from the users based on the shopping-related index of each user.

[0077] In this application, shopping-related indicators such as browsing rate, conversion rate, and brand penetration rate can be analyzed for each user who meets the target selection criteria. Among them, the browsing rate is used to measure the attention of users who meet the target selection criteria to the product category, and the specific calculation method is: the number of users who have browsed the blood glucose meter among all users of the target selection criteria ÷ all users of the target selection criteria × 100%. The conversion rate is used to evaluate the conversion rate from browsing to purchase of users who meet the target selection criteria, and the specific calculation method is: the number of users who have browsed and purchased blood glucose meters among all users of the target selection criteria ÷ the number of users who have browsed blood glucose meters among all users of the target selection criteria × 100%. The brand penetration rate is used to reflect the brand awareness and purchase intention of users who meet the target selection criteria, and the specific calculation method is: the number of users who have purchased blood glucose meters of this brand among all users of the target selection criteria ÷ the number of users who have browsed and purchased blood glucose meters among all users of the target selection criteria × 100%.

[0078] In order to further classify the users of the target selection criteria and locate potential users more accurately, we calculate the benchmark browsing rate: the number of all users who have browsed blood glucose meters ÷ all users × 100%, the benchmark conversion rate: the number of all users who have browsed and purchased blood glucose meters ÷ the number of all users who have browsed and purchased blood glucose meters × 100%, and the benchmark brand penetration rate: the number of all users who have purchased blood glucose meters of this brand ÷ the number of all users who have browsed and purchased blood glucose meters × 100%.

[0079] Specifically, the possibility of converting users with target selection conditions of browsing rate lower than the benchmark browsing rate, conversion rate lower than the benchmark conversion rate, and brand penetration rate lower than the benchmark brand penetration rate into brand users is the lowest, so this user is regarded as the brand elimination population. The possibility of converting users with target selection conditions of browsing rate higher than the benchmark browsing rate, conversion rate higher than the benchmark conversion rate, or brand penetration rate higher than the benchmark brand penetration rate into brand users is high, so this user is regarded as the brand browsing population. The possibility of converting users with target selection conditions of browsing rate lower than the benchmark browsing rate, conversion rate higher than the benchmark conversion rate, or brand penetration rate higher than the benchmark brand penetration rate into brand users is very high, so this user is regarded as the brand interest population. For example: users of "obese people" on e-commerce platform B are stratified into the brand elimination population, users of "gift-giving people" are stratified into the brand browsing population, and users of "sweet lovers" are stratified into the brand interest population.

[0080] Brand browsing groups and brand interest groups that meet the target selection criteria can be identified as potential users of the target product. For example, users who are "gift-giving groups" and users who are "sweet tooth lovers" can be identified as potential users of the target product.

[0081] At the same time, when the user's query, browsing, collection, purchase and other behavior data can be obtained, the user who shows a high interest in a certain product or a certain category of products, i.e., potential users, can be identified by analyzing the user's behavior data. For example, users who frequently browse or collect a certain category of products may be potential buyers of the product. Then, the potential customers determined by the behavior data and the potential customers determined by the text can be merged, etc. The specific merging method can be simple merging, weighted merging or intersection merging.

[0082] Specifically, simple union refers to directly merging the potential customer lists obtained based on behavioral data and text analysis to form a larger potential customer group. This method is simple and easy to implement, but may contain duplicate or less relevant users. Weighted merging refers to assigning different weights to potential customers from different sources. For example, when behavioral data is more reliable than text analysis, potential customers determined based on behavioral data can be given a higher weight. Intersection merging refers to finding users who appear in the potential customer lists obtained based on both behavioral data and text analysis. These users may have a stronger interest and demand for specific products or categories and are therefore more valuable potential customers. In practical applications, the way to merge potential customers should be selected according to specific business scenarios and goals. For example, if the goal is to improve the accuracy of marketing activities, taking the intersection may be a better choice; if the goal is to expand brand influence, simple union may be more appropriate.

[0083] When testing based on potential users, if the number of potential users is not enough to meet the test requirements, you can consider relaxing the screening conditions appropriately to expand the scale of users. For example, on e-commerce platform A, there is usually no problem of exposure and difficulty in increasing the number of people in a population package of more than 500,000. For target selection conditions with a population package of less than 500,000, the search behavior of the target selection condition in the past 30 days can be adjusted to the search behavior in the past 60 days, thereby increasing the number of people in the population package.

[0084] In summary, the potential user identification method provided by the present application can obtain text crawled using keywords of commodities, and perform semantic analysis on the text to obtain description tags, further screen the description tags, and obtain matching target selection conditions using target description tags, thereby identifying potential users. The present application can use text to identify potential users, solving the problem of the prior art that potential users cannot be accurately identified when relevant data on user query, browsing, collection, and purchase behaviors are missing.

[0085] Furthermore, after selecting the target description tag from the description tags according to the number of texts with each description tag, the method further includes: verifying the target description tag by using multi-channel information.

[0086] It should be noted that user reviews, texts left on social platforms and other texts often reflect users' subjective cognition and needs, but whether these cognitions and needs are truly valid in the medical field needs to be verified based on scientific research and professional literature.

[0087] In this application, desktop research can be used to verify the target description tag, that is, text analysis and data verification technology can be used to deeply explore the information sources of existing literature, network resources and professional forums, and further verify the user's stated needs. At this time, the target selection conditions that match the verified target description tag are screened out from the preset user selection condition library.

[0088] Specifically, cross-validation: by comparing information from different information sources, such as discussions in medical forums, research results in professional literature, and descriptions in user reviews, to ensure that the identified needs are real and reliable. Eliminate pseudo-needs: During the verification process, if it is found that some users’ claimed needs lack scientific basis or are inconsistent with the actual situation, these needs will be regarded as pseudo-needs and eliminated.

[0089] To sum up, the method for verifying target description labels using multi-channel information provided in this application can ensure that these labels not only reflect the user's subjective cognition and needs, but are also based on scientific research and professional knowledge, thereby improving their scientificity and objectivity.

[0090] Furthermore, after determining at least some of the users who meet the target selection criteria as potential users of the target product, the method further includes: sending a test advertisement to at least some of the potential users to obtain a test result.

[0091] In this application, A / B testing can be used to compare the responses of users with different target selection criteria to the same advertisement, thereby finding the user with the highest responsiveness. The target selection criteria can also be adjusted based on the test data results to retain valuable users.

[0092] Specifically, a detailed test plan is developed, including the test target, the test population, the test ad content, the test time, etc. Multiple shopping platforms are selected for testing to obtain more extensive data. In order to accurately evaluate the response of different groups to the advertisement, factors such as the reach channel, delivery time, and advertising materials are used as control variables to ensure that these factors remain consistent during the test process. Then the test ads are placed on the advertising space of the shopping platform to ensure that users with different target selection conditions see the same test ads in the same location and at the same time, and the effect data of the test ads are collected, including click-through rate, collection and purchase rate, conversion rate, input-output ratio, etc. In order to avoid large fluctuations in data when the number of clicks is too small, resulting in the inability to accurately reflect the real response of the population, and the possibility of not obtaining enough subsequent collections / add-to-carts / purchases, a minimum value is set, that is, the number of clicks for each population package must reach at least fifty times. Analyze the collected data to find the population package with the highest responsiveness to optimize the positioning of potential users of the target product.

[0093] For ease of understanding, here are some examples: The input-output ratio of users who are "sweet tooth lovers" is lower than the average level, but the test population is large and can be optimized before retesting. Therefore, the target selection conditions of first-tier and second-tier cities can be superimposed and retested. The input-output ratio of users who are "fat loss people" can reach more than 18, indicating that the conversion potential is high. There is still room to expand the population size in the future to obtain greater exposure and conversion. Therefore, the conditions of browsing blood glucose meters can be eliminated, and high consumption power, first-tier cities and second-tier cities can be superimposed for retesting.

[0094] In summary, the method of sending test advertisements to potential users provided in the present application can accurately identify the user group with the highest responsiveness by testing and comparing the reactions of users with different target selection conditions to the same advertisement, thereby more accurately locating the potential users of the target product. This helps merchants to focus their limited marketing resources on the most valuable users. The target selection conditions can also be adjusted according to the test data results to eliminate user groups that do not respond well to the advertisements and retain or increase user groups that respond positively to the advertisements. This dynamic adjustment helps merchants to continuously optimize the positioning of potential users and improve marketing efficiency.

[0095] A potential user identification method provided by an embodiment of the present application is introduced above, and a device for executing the potential user identification method will be introduced below.

[0096] See also Figure 2 , Figure 2 This is a schematic diagram of the structure of a potential user identification device provided in an embodiment of the present application. Figure 3 As shown, the potential user identification device comprises:

[0097] The obtaining unit 11 is used to obtain a plurality of texts crawled from the network according to a target keyword, where the target keyword is a keyword of a target product.

[0098] The analysis unit 12 is used to perform semantic analysis on part or all of the content in each of the texts, and add description tags to each of the texts based on the semantic analysis results.

[0099] The first screening unit 13 is used to screen out a target description tag from each of the description tags according to the number of the texts with each of the description tags.

[0100] The second screening unit 14 is used to screen out target selection conditions that match the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting users in a shopping platform.

[0101] The determination unit 15 is configured to determine at least some of the users who meet the target selection condition as potential users of the target product.

[0102] In a possible implementation, the analysis unit 12 for performing semantic analysis on part or all of the content in each of the texts includes:

[0103] The word segmentation subunit is used to segment each of the texts to obtain a plurality of words, and filter out crowd type related words and shopping demand related words from the plurality of words.

[0104] The construction subunit is used to construct a phrase containing crowd type-related words and / or shopping demand-related words, wherein the phrase includes a verb.

[0105] The first analysis subunit is used to perform semantic analysis on the phrase.

[0106] In a possible implementation, the analysis unit 12 for performing semantic analysis on part or all of the content in each of the texts includes:

[0107] The identification subunit is used to use a preset rule matching algorithm to identify phrases in the text that meet the preset rule matching algorithm.

[0108] The second analysis subunit is used to perform semantic analysis on the phrase.

[0109] In a possible implementation, the device further includes:

[0110] A verification unit is used to verify the target description tag using multi-channel information.

[0111] The second screening unit 14 comprises:

[0112] The first screening subunit is used to screen out target selection conditions that match the verified target description tag from a preset user selection condition library.

[0113] In a possible implementation, the determining unit 15 includes:

[0114] The prediction subunit is used to predict at least one shopping-related indicator of each user who meets the target selection condition.

[0115] The second screening subunit is used to screen out potential users of the target product from the users according to the shopping association index of each user.

[0116] In a possible implementation, the device further includes:

[0117] The sending unit is used to send the test advertisement to at least some of the potential users to obtain the test result.

[0118] The present application also provides an electronic device in an embodiment. Figure 3 As shown, it shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiment of the present application. The electronic device in the embodiment of the present application may include but is not limited to fixed terminals such as mobile phones, laptops, PDAs (personal digital assistants), PADs (tablet computers), desktop computers, etc. Figure 3 The electronic device shown is merely an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0119] like Figure 3 As shown, the electronic device may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 301, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 302 or a program loaded from a storage device 308 to a random access memory (RAM) 303. When the electronic device is powered on, various programs and data required for the operation of the electronic device are also stored in the RAM 303. The processing device 301, the ROM 302, and the RAM 303 are connected to each other via a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0120] Typically, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touch pad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a memory card, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device to communicate with other devices wirelessly or by wire to exchange data. Although Figure 3An electronic device having various devices is shown, but it should be understood that it is not required to implement or possess all the devices shown. More or fewer devices may be implemented or possessed instead.

[0121] An embodiment of the present application also provides a computer program product including computer-readable instructions. When the computer-readable instructions are executed on an electronic device, the electronic device implements any potential user identification method provided in the embodiment of the present application.

[0122] A computer-readable storage medium is also provided in an embodiment of the present application. The storage medium carries one or more computer programs. When the one or more computer programs are executed by an electronic device, the electronic device can implement any potential user identification method provided in the embodiment of the present application.

[0123] It should also be noted that the device embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed over multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. In addition, in the drawings of the device embodiments provided by the present application, the connection relationship between the modules indicates that there is a communication connection between them, which may be specifically implemented as one or more communication buses or signal lines.

[0124] Through the description of the above implementation mode, the technicians in the field can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course, it can also be implemented by special hardware including special integrated circuits, special CPUs, special memories, special components, etc. In general, all functions completed by computer programs can be easily implemented by corresponding hardware, and the specific hardware structure used to implement the same function can also be various, such as analog circuits, digital circuits or special circuits. However, for the present application, software program implementation is a better implementation mode in more cases. Based on such an understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer floppy disk, a U disk, a mobile hard disk, a ROM, a RAM, a disk or an optical disk, etc., including a number of instructions to enable a computer device (which can be a personal computer, a training device, or a network device, etc.) to execute the methods described in each embodiment of the present application.

[0125] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0126] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from a website site, a computer, a training device, or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, training device, or data center. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium may be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)), etc.

Claims

1. A potential user identification method, characterized in that: include: Obtaining multiple texts crawled from the network according to target keywords, wherein the target keywords are keywords of target products; Performing semantic analysis on part or all of the content in each of the texts, and adding description tags to each of the texts based on the semantic analysis results; Filtering out a target description tag from the description tags according to the number of the texts with the description tags; Filtering out a target selection condition that matches the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting a user on a shopping platform; At least some of the users who meet the target selection conditions are determined as potential users of the target product.

2. The potential user identification method according to claim 1, characterized in that: The semantic analysis of part or all of the content in each of the texts includes: Segmenting each of the texts to obtain a plurality of words, and screening out crowd type-related words and shopping demand-related words from the plurality of words; Constructing a phrase including vocabulary associated with a crowd type and / or vocabulary associated with a shopping demand, wherein the phrase includes a verb; The phrase is semantically analyzed.

3. The potential user identification method according to claim 1, characterized in that: The semantic analysis of part or all of the content in each of the texts includes: Using a preset rule matching algorithm, identifying phrases in the text that conform to the preset rule matching algorithm; The phrase is semantically analyzed.

4. The potential user identification method according to claim 1, characterized in that: After filtering out a target description tag from each of the description tags according to the number of the texts with each of the description tags, the method further includes: Verifying the target description label using multi-channel information; The step of selecting a target selection condition that matches the target description tag from a preset user selection condition library includes: The target selection condition that matches the verified target description tag is screened out from the preset user selection condition library.

5. The potential user identification method according to claim 1, characterized in that: The step of determining at least some of the users who meet the target selection condition as potential users of the target product includes: Predicting at least one shopping-related indicator for each user who meets the target selection condition; Potential users of the target product are screened out from the users according to the shopping association index of each user.

6. The potential user identification method according to claim 1, characterized in that: After determining at least some of the users who meet the target selection condition as potential users of the target product, the method further includes: Send a test advertisement to at least some of the potential users to obtain a test result.

7. A potential user identification method and device, characterized in that: include: An obtaining unit, used to obtain a plurality of texts crawled from the network according to a target keyword, wherein the target keyword is a keyword of a target product; An analysis unit, configured to perform semantic analysis on part or all of the content in each of the texts, and to add a description tag to each of the texts based on the semantic analysis result; A first screening unit, configured to screen out a target description tag from each of the description tags according to the number of the texts with each of the description tags; A second screening unit, configured to screen out a target selection condition that matches the target description tag from a preset user selection condition library, wherein the preset user selection condition library includes a plurality of conditions for selecting a user on a shopping platform; The determination unit is used to determine at least some of the users who meet the target selection conditions as potential users of the target product.

8. A computer program product, characterized in that The method comprises computer-readable instructions, which, when executed on an electronic device, enable the electronic device to implement the potential user identification method as claimed in any one of claims 1 to 6.

9. An electronic device, characterized in that: The method comprises at least one processor and a memory connected to the processor, wherein: The memory is used to store computer programs; The processor is used to execute the computer program so that the electronic device can implement the potential user identification method as described in any one of claims 1 to 6.

10. A computer storage medium, characterized in that: The storage medium carries one or more computer programs, and when the one or more computer programs are executed by an electronic device, the electronic device can implement the potential user identification method as described in any one of claims 1 to 6.

Images