Search engine based user portrait construction method, system, medium and device

Abstract

A search engine-based user portrait construction method, system, medium and device, relating to the technical field of Internet, the method comprises: obtaining target search behavior data of a user; determining data corresponding to static feature labels and dynamic feature labels of the target user according to the target search behavior data; calculating first feature weights of the user according to the data corresponding to each dynamic feature label; correcting the first feature weights based on a preset time decay formula to obtain second feature weights; and generating a user portrait of the user based on each static feature label and the second feature weights. The technical solution provided by the present application can construct a more accurate user portrait according to the search behavior of the user, thereby facilitating subsequent recommendation of more accurate related information to the user based on the user portrait.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of Internet, in particular to a user portrait construction method and system based on a search engine, a medium and equipment. BACKGROUND

[0002] A user portrait is a common and effective tool for sketching target users, contacting user demands and designing directions. The user portrait has been widely used in various fields, especially in the Internet field. The user role formed by the user portrait is not constructed outside the product and market, but a virtual role representing the main audience and target group of the product.

[0003] The traditional user portrait is mainly constructed by common and widely used user characteristics. The user portrait may be inaccurate due to the reasons such as the closed search related data source and the non-centralized user search characteristics, thereby causing the inaccurate related information recommended for the user. SUMMARY

[0004] The present application provides a user portrait construction method and system based on a search engine, a medium and equipment, which can construct a more accurate user portrait according to the search behavior of the user, thereby facilitating the subsequent recommendation of more accurate related information for the user according to the user portrait.

[0005] In a first aspect, the present application provides a user portrait construction method based on a search engine, which comprises:

[0006] Obtaining target search behavior data of a user;

[0007] Determining data corresponding to static characteristic labels and dynamic characteristic labels of the target user according to the target search behavior data; calculating a first characteristic weight of the user according to the data corresponding to each dynamic characteristic label;

[0008] Correcting the first characteristic weight based on a preset time decay formula to obtain a second characteristic weight;

[0009] Generating a user portrait of the user based on each static characteristic label and the second characteristic weight.

[0010] By adopting the above technical solution, the static feature tags and dynamic feature tags of the target user can be determined based on the user's target search behavior data. The user's first feature weight can be calculated based on the data corresponding to the dynamic feature tags. Since the user's behavior data may change over time, the first feature weight can be corrected based on a preset time decay formula to obtain the second feature weight. Then, a user profile is generated based on the static feature tags and the second feature weight, so that the user profile can be dynamically updated with the user's target search behavior, which improves the accuracy of user profile construction and facilitates the subsequent recommendation of more accurate relevant information to the user based on the user profile.

[0011] Optionally, obtaining the user's target search behavior data includes: obtaining all of the user's search behavior data; cleaning the search behavior data to obtain cleaned search behavior data; calculating the correlation between each data in the cleaned search behavior data; and filtering out data with a correlation higher than a correlation threshold to obtain the target search behavior data.

[0012] By adopting the above technical solution, all user search behavior data can be cleaned to remove useless and abnormal data, improve data accuracy, calculate the correlation between cleaned data, and filter out data with correlation values ​​higher than the correlation threshold, further improving the accuracy of target search behavior data.

[0013] Optionally, determining the user's static feature label based on the target search behavior data includes: obtaining keywords from the target search behavior data; matching the keywords with a preset keyword library, and taking the keyword with the highest matching degree as the target keyword; inputting the target keyword into a trained user attribute prediction model to obtain multiple predicted user attributes, and taking the user attributes as the static feature label.

[0014] By adopting the above technical solution, since there may be cases where user attributes are missing or inaccurate, predictions can be made based on keywords in the user's target search behavior data to obtain the user's static feature tags. This allows for the subsequent construction of user profiles based on static feature tags, further improving the accuracy of user profiles.

[0015] Optionally, the first feature weight of the user is calculated based on data corresponding to each of the dynamic feature tags, comprising: filtering a search frequency tag, a search path tag and a search click tag from each of the dynamic feature tags; calculating a weight of the search frequency tag, a weight of the search path tag and a weight of the search click tag respectively; multiplying the weight of the search frequency tag, the weight of the search path tag and the weight of the search click tag to obtain the first feature weight.

[0016] By using the above technical solution, the weight of the search frequency tag, the weight of the search path tag and the weight of the search click tag are calculated, and the first feature weight is obtained by multiplying the weights of the three tags. The first feature weight of the comprehensive dynamic feature tag can be obtained by comprehensively calculating the three representative dynamic feature tags.

[0017] Optionally, the first feature weight is corrected based on a preset time decay formula to obtain a second feature weight, comprising: obtaining a current date and a previous date of the last search behavior data of the target user; substituting the current date, the previous date and the first feature weight into the preset time decay formula to obtain the second feature weight; the preset time decay formula is Y=X*exp[-α(d1-d2)], wherein Y is the second feature weight, X is the first feature weight, α is an attenuation factor, d1 is the current date, and d2 is the previous date.

[0018] By using the above technical solution, the first feature weight can be iteratively corrected according to the search behavior of the user and the time factor, and the accuracy of the user portrait is further improved.

[0019] Optionally, the user portrait of the user is generated based on each of the static feature tags and the second feature weight, comprising: determining a user level according to the second feature weight and a weight-level correspondence relationship; and adding the user level and the static feature tags to the user portrait of the user.

[0020] By using the above technical solution, the user level can be determined according to the second feature weight, and the user portrait is generated according to the user level and the static feature tags, thereby improving the accuracy of the user portrait.

[0021] Optionally, after the user portrait of the user is generated based on the static feature labels and the second feature weights, the method further includes: determining related information of interest to the user based on the user portrait; matching a corresponding information recommendation template according to the user level, and matching the related information to the information recommendation template to obtain target recommendation information, the information template including at least one dimension of a related industry, a risk warning, an added enterprise, an added patent, and news updates; and pushing the target recommendation information to the user.

[0022] By using the above technical solution, the related information of interest to the user is determined according to the user portrait, and a corresponding information recommendation template is matched according to the user level, for example, some relatively simple and understandable information can be recommended to a primary user. The corresponding recommendation template can be matched for the user according to the user level, thereby improving the experience of the user.

[0023] In a second aspect of the present application, a user portrait construction system based on a search engine is provided, and the system includes:

[0024] A search behavior data acquisition module is configured to acquire target search behavior data of a user.

[0025] A label determination module is configured to determine static feature labels and data corresponding to dynamic feature labels of the target user according to the target search behavior data.

[0026] A first feature weight calculation module is configured to calculate first feature weights of the user according to the data corresponding to the dynamic feature labels.

[0027] A second feature weight correction module is configured to correct the first feature weights based on a preset time decay formula to obtain second feature weights.

[0028] A user portrait generation module is configured to generate a user portrait of the user based on the static feature labels and the second feature weights.

[0029] In a third aspect of the present application, a computer storage medium is provided, and the computer storage medium stores a plurality of instructions, the instructions being suitable for being loaded by a processor and performing the method steps described above.

[0030] In a fourth aspect of the present application, an electronic device is provided, and the electronic device includes a processor and a memory, wherein the memory stores a computer program, and the computer program is suitable for being loaded by the processor and performing the method steps described above.

[0031] In summary, one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages:

[0032] 1、The application can determine the data corresponding to the static feature label and the dynamic feature label of the target user according to the target search behavior data of the user, and can calculate the first feature weight of the user based on the data corresponding to the dynamic feature label. Since the behavior data of the user may change over time, the first feature weight is corrected based on the preset time decay formula to obtain the second feature weight, and then the user portrait is generated based on the static feature label and the second feature weight, so that the user portrait can be dynamically updated with the target search behavior of the user, improving the accuracy of user portrait construction, so as to facilitate subsequent recommendation of more accurate related information to the user according to the user portrait.

[0033] 2、The application can determine the related information interested by the user according to the user portrait, and match the corresponding information recommendation template according to the user level, such as recommending some relatively simple and understandable information to the primary user. The corresponding recommendation template can be matched for the user according to the user level, improving the experience of the user. BRIEF DESCRIPTION OF DRAWINGS

[0034] Figure 1 is a flow diagram of a user portrait construction method based on a search engine provided by an embodiment of the application;

[0035] Figure 2 is a module diagram of a user portrait construction system based on a search engine provided by an embodiment of the application;

[0036] Figure 3 is a structural diagram of an electronic device provided by an embodiment of the application.

[0037] BRIEF DESCRIPTION OF DRAWINGS: 1, search behavior data acquisition module; 2, label determination module; 3, first feature weight calculation module; 4, second feature weight correction module; 5, user portrait generation module; 300, electronic device; 301, processor; 302, communication bus; 303, user interface; 304, network interface; 305, memory. DETAILED DESCRIPTION

[0038] In order for those skilled in the art to better understand the technical solutions in the specification, the technical solutions in the specification will be described clearly and completely in conjunction with the drawings in the embodiments of the specification. Obviously, the described embodiments are only part of the embodiments of the application, not all embodiments.

[0039] In the description of the embodiments of the present application, the words such as "for example" or "for instance" are used to represent an example, an illustration or an exposition. Any embodiment or design scheme described as "for example" or "for instance" in the embodiments of the present application should not be interpreted as more preferred or more advantageous than other embodiments or design schemes. In fact, the words such as "for example" or "for instance" are intended to present the relevant concept in a specific manner.

[0040] In the description of the embodiments of the present application, the term "a plurality of" means two or more. For example, a plurality of systems means two or more systems, and a plurality of screen terminals means two or more screen terminals. In addition, the terms "first" and "second" are used only for the purpose of description, and should not be interpreted as indicating or implying relative importance or implicitly indicating the indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more features. The terms "include", "contain", "have" and their variants mean "include but are not limited to", unless otherwise specifically emphasized.

[0041] In one embodiment, please refer to Figure 1 , a flowchart of a user portrait construction method based on a search engine is specially proposed. The method can be implemented by relying on a computer program, can be implemented by relying on a single-chip microcomputer, and can run on a system for constructing a user portrait based on a search engine. The computer program can be integrated in an application or run as an independent tool application. The method includes steps 10 to 50, and the above steps are as follows:

[0042] Step 10: Obtain target search behavior data of a user.

[0043] With the development of network technology, more and more services are provided through the network, and users can search on the network platform to obtain various information. The target search behavior data in the embodiments of the present application can be a search behavior record of a user in a characteristic time range.

[0044] On the basis of the above-mentioned embodiments, as an optional embodiment, the step of obtaining target search behavior data of a user can include the following steps:

[0045] Step 101: Obtain all search behavior data of a user.

[0046] Specifically, a user can obtain a product use user and related search behavior data on a shenshi platform,

[0047] Step 102: Perform data cleaning on the search behavior data to obtain cleaned search behavior data.

[0048] Specifically, since there are a large number of search behavior data of the user, some of the behavior data are abnormal or invalid, and these behavior data have no actual meaning, but will increase the subsequent calculation amount. Therefore, after obtaining all the search behavior data of the user, the search behavior data needs to be accurately cleaned, and a certain cleaning standard can be set, such as removing duplicate data, removing abnormal data, and the like, and the data that does not meet the cleaning standard is removed to obtain the cleaned search behavior data, which can ensure the quality and usability of the data.

[0049] Step 103: calculating the correlation between each data in the cleaned search behavior data.

[0050] Step 104: screening out data with a correlation higher than a correlation threshold to obtain target search behavior data.

[0051] Specifically, since the user may search with a behavior target in the search process, it may be based on a product or a search target such as an enterprise, but some irrelevant content to the search purpose may be searched in the search process, or there are searches performed by others under the user. Therefore, the correlation between each data in the cleaned search behavior data needs to be calculated, and the specific correlation calculation method can be a correlation coefficient calculation method or a factor analysis method, which is not described herein. The data with a correlation higher than a correlation threshold can be screened out to obtain target search behavior data. That is, all search behavior data close to the user's search target can be screened out to obtain more accurate target search behavior data.

[0052] Step 20: determining the data corresponding to the static feature label and the dynamic feature label of the target user according to the target search behavior data.

[0053] The static feature label in the embodiment of the application can be some basic information of the user, which classifies and labels the user, such as the user's gender, age, education, occupation, geographical location, and the like. The dynamic feature label in the embodiment of the application refers to a way of classifying and labeling the user based on the user's real-time behavior and interest information, such as the user's search frequency, frequently accessed search page, search time point, search word average length, and search click. The data corresponding to the dynamic feature label refers to the data value corresponding to each dynamic label.

[0054] On the basis of the above embodiment, as an optional embodiment, the step of determining the static feature label of the target user according to the target search behavior data can include the following steps:

[0055] Step 201: obtaining the keywords of the target search behavior data.

[0056] Step 202: match the keyword with the preset keyword library, and take the keyword with the highest matching degree as the target keyword

[0057] Step 203: input the target keyword into the trained user attribute prediction model to obtain the predicted multiple user attributes, and take the user attributes as the static feature label.

[0058] Specifically, since the personnel may log in before searching and need to fill in personal basic information, but often due to the user privacy, the user may choose not to fill in or fill in the wrong information, or the user directly does not log in and searches, at this time, the static feature label of the user needs to be predicted to establish the real attribute of the user. The keyword filtering rule can be set to identify the keyword in the target search behavior data of the user, match the keyword with the preset keyword library, the preset keyword library can include popular keywords, competitor keywords, etc., and classify and sort according to similarity, frequency and the like. The matching degree of each keyword and each keyword in the preset keyword library is calculated, and the keyword with the highest matching degree is taken as the target keyword. The matching degree can be calculated by edit distance, cosine similarity and the like.

[0059] Further, before obtaining the keyword of the target search behavior data, an initial user attribute prediction model is constructed, and the keyword sample is selected to train the initial user attribute prediction model until the training is stopped when the training is completed to the training completion standard, that is, the prediction accuracy of the user attribute prediction model reaches a certain preset value, such as 90%. The target keyword is input into the trained user attribute prediction model, and the predicted multiple user attributes are output, and the user attributes are taken as the static feature label.

[0060] Step 30: calculate the first feature weight of the user according to the data corresponding to each dynamic feature label.

[0061] Based on the above embodiment, as an optional embodiment, the step of calculating the first feature weight of the user according to the data corresponding to each dynamic feature label can further include the following steps:

[0062] Step 301: screen out the search frequency label, search path label and search click label from each dynamic feature label.

[0063] Specifically, the search frequency label, the search path label and the search click label are screened out from the dynamic feature labels. The search frequency refers to the frequency of the user accessing the search page or using the search function, which can represent the search frequency habit of the user. The search path refers to the page / path accessed by the user, which can represent the search willingness degree of the user. The search click refers to the click amount of the user from sending a search request to the search result, which can represent whether the search demand of the user is met.

[0064] Step 302: The weight of the search frequency label, the weight of the search path label and the weight of the search click label are calculated respectively.

[0065] Step 303: The weight of the search frequency label, the weight of the search path label and the weight of the search click label are multiplied to obtain the first feature weight.

[0066] Specifically, the data corresponding to the search frequency label, the data corresponding to the search path label and the data corresponding to the search click label are obtained respectively. The weight of the search frequency label = search times / total search times, the search times refer to the number of times that the user searches for a certain search target / keyword within a period of time, and the total search times refer to the total number of times that the user searches for all search targets / all keywords within a period of time. The weight of the search path label = manual input search click times / total click times, the manual input search click times refer to the number of times that the user manually inputs a keyword to search, and the total click times refer to the total number of clicks of the user searching through all paths, including the number of times that the user manually inputs the search click times and the number of times that the user directly clicks various label hot searches on the search interface. The weight of the search click label = search result interface click times / total click times, the search result interface click times refer to the number of times that the user clicks a certain result related to the search keyword on the search result interface after searching, and the total click times refer to the total number of clicks of the user on the search result page. The weight of the search frequency label, the weight of the search path label and the weight of the search click label are multiplied to obtain the first feature weight, which can represent the comprehensive weight of the user in the dynamic label.

[0067] It should be noted that in a feasible embodiment, the weights corresponding to the data of other dynamic feature labels can also be calculated and multiplied to obtain the first feature weight, which is not limited herein.

[0068] Step 40: The first feature weight is corrected based on the preset time decay formula to obtain the second feature weight.

[0069] Specifically, as the search behavior data of the user changes over time, the user may be more interested in product A in the early period and more interested in product B in the later period, so that the interest in product A decreases. Therefore, in order to further improve the accuracy of the user portrait, the first feature weight is corrected based on a preset time decay formula to obtain a second feature weight. The time decay formula is Y=X*exp[-a(d1-d2)], where Y is the second feature weight, X is the first feature weight, a is the decay factor, which can be adjusted and set according to actual conditions, d1 is the current date, and d2 is the previous date. By correcting the first feature weight through the time decay formula, the second feature weight can be kept timely, and the accuracy of the user portrait is further improved.

[0070] Step 50: generating a user portrait of the user based on the static feature labels and the second feature weight.

[0071] Specifically, in the embodiment of the present application, there is an initial user portrait for each user. After obtaining the static feature labels and the second feature weight of the user, the second feature weight is determined to fall into which weight interval corresponding to the level according to the second feature weight and the weight-level correspondence relationship, and then the corresponding user level can be obtained. The greater the weight, the higher the corresponding level, and the higher the level, the better the user portrait. Subsequently, information can be recommended to the user according to the user level in the user portrait. The static feature labels and the second feature weight are added to the initial user portrait, and the user portrait corresponding to the user can be generated.

[0072] On the basis of the above embodiment, as an optional embodiment, after the step of generating a user portrait of the user based on the static feature labels and the second feature weight, a step of recommending information to the user according to the user portrait can be further included. The step can specifically include:

[0073] Since the static feature label of the user and the second feature weight of a certain search behavior are included in each user portrait, the relevant information of interest to the user can be determined according to the user portrait. If the user portrait shows that the user is a technology enthusiast, information related to technology can be recommended, such as the latest technology news, innovative technology products, etc. Since each user has a different level, the higher the level, the more the user is interested in the corresponding information recommendation template according to the user level. Different levels of users may be interested in different types of information, so the corresponding information recommendation template needs to be selected according to the user level. For example, for a primary user, some simple and easy-to-understand information can be recommended, and for a senior user, more in-depth and professional information can be recommended. The information recommendation template can include dimensions such as related industries, risk warnings, new enterprises, new patents, and news updates. By matching the relevant information with the information template, the target recommendation information that matches the user's interest and demand can be obtained. Finally, the target recommendation information is pushed to the user through various ways. The form and frequency of the push can be adjusted according to the user's preferences and behavior habits to improve the effect of the push and the user experience.

[0074] In another possible embodiment, based on the user portrait, the search behavior data of each user can be analyzed irregularly, each user can be divided into clusters, the characteristics and trends of the search behavior data of the users in each cluster can be analyzed, and through data mining and analysis, the user's demand and preference can be obtained, and adjustment suggestions for subsequent search product iteration and operation strategy can be provided, such as optimizing the search algorithm, improving the search result page layout, recommending related content, and optimizing the advertising placement strategy. Suggestions need to be combined with specific business scenarios and user needs to ensure the feasibility and effectiveness of the implementation plan. After the search product iteration and operation strategy adjustment, data can still be tracked and analyzed to evaluate the effect and improve the scheme. Data monitoring and analysis tools can be used to regularly collect and analyze search behavior data, and feedback results and suggestions can be provided to relevant teams and decision makers to provide data support and basis for business decisions.

[0075] See Figure 3 A module schematic diagram of a user portrait construction system based on a search engine provided by the embodiment of the application can include: a search behavior data acquisition module 1, a label determination module 2, a first feature weight calculation module 3, a second feature weight correction module 4, and a user portrait generation module 5.

[0076] The search behavior data acquisition module 1 is used to acquire the target search behavior data of the user.

[0077] The label determination module 2 is used to determine the data corresponding to the static feature label and the dynamic feature label of the target user according to the target search behavior data.

[0078] The first feature weight calculation module 3 is configured to calculate the first feature weight of the user according to data corresponding to each dynamic feature label.

[0079] The second feature weight correction module 4 is configured to correct the first feature weight based on a preset time decay formula to obtain a second feature weight.

[0080] The user portrait generation module 5 is configured to generate a user portrait of the user based on each static feature label and the second feature weight.

[0081] It should be noted that the system provided in the above embodiments only divides the above-mentioned function modules for example to achieve its functions. In actual application, the above-mentioned functions can be completed by different function modules according to needs, that is, the internal structure of the device is divided into different function modules to complete all or part of the above-described functions. In addition, the system and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process is described in the method embodiment, which will not be described here.

[0082] The present application also discloses an electronic device. Referring to Figure 3 , Figure 3 is a structural schematic diagram of an electronic device disclosed by the embodiments of the present application. The electronic device 300 can include at least one processor 301, at least one network interface 304, a user interface 303, a memory 305, and at least one communication bus 302.

[0083] The communication bus 302 is configured to realize the connection and communication between the components.

[0084] The user interface 303 can include a display screen (Display) and a camera (Camera). Optionally, the user interface 303 can further include a standard wired interface and a wireless interface.

[0085] The network interface 304 can optionally include a standard wired interface and a wireless interface (such as a WI-FI interface).

[0086] The processor 301 can include one or more processing cores. The processor 301 connects various parts within the server through various interfaces and lines, performs various functions of the server and processes data by running or executing instructions, programs, code sets or instruction sets stored in the memory 305, and calling data stored in the memory 305. Alternatively, the processor 301 can be implemented in at least one of a hardware form of a digital signal processing (DSP), a field-programmable gate array (FPGA), and a programmable logic array (PLA). The processor 301 can integrate a combination of one or more of a central processing unit (CPU), a graphics processing unit (GPU), and a modem. Among them, the CPU mainly processes operating systems, user interfaces, and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; and the modem is used for processing wireless communication. It can be understood that the above-mentioned modem can also not be integrated into the processor 301, but can be realized by a separate chip.

[0087] The memory 305 can include a random access memory (RAM) and a read-only memory (ROM). Alternatively, the memory 305 includes a non-transitory computer-readable storage medium. The memory 305 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 305 can include a program storage area and a data storage area, wherein the program storage area can store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playing function, an image playing function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area can store data involved in the above-mentioned various method embodiments, etc. The memory 305 can alternatively be at least one storage device located away from the aforementioned processor 301. Referring to Figure 3 The memory 305 as a computer storage medium can include an operating system, a network communication module, a user interface module, and an application program of a user portrait construction method based on a search engine.

[0088] In Figure 3In the electronic device 300 shown, the user interface 303 is mainly used to provide an interface for the user to input, and obtain data input by the user; and the processor 301 can be used to invoke an application program stored in the memory 305 and storing a search engine-based user portrait construction method, which, when executed by one or more processors 301, causes the electronic device 300 to perform the method described in one or more of the above embodiments. It should be noted that, for the foregoing method embodiments, in order to simply describe, they are all described as a combination of a series of actions, but those skilled in the art should know that the present application is not limited to the order of the actions described, because according to the present application, certain steps can be performed in other order or at the same time. Secondly, those skilled in the art should know that the embodiments described in the specification all belong to preferred embodiments, and the actions and modules involved are not necessarily required by the present application.

[0089] In the above embodiments, the description of each embodiment has its own focus, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments.

[0090] In the several embodiments provided by the present application, it should be understood that the disclosed device can be implemented in other ways. For example, the device embodiments described above are only schematic. The division of the units is only a logical function division. There can be another division manner for actual implementation, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections between the units can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical or other forms.

[0091] The units described as separate components can or can not be physically separate, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the embodiment.

[0092] In addition, each functional unit in each embodiment of the present application can be integrated into a processing unit, or each unit can exist physically independently, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0093] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable memory. Based on such understanding, the technical solutions of the present application essentially or the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a memory and includes a plurality of instructions for causing a computer device (which can be a personal computer, a server or a network device, etc.) to execute all or part of the steps of the embodiments of the present application. The aforementioned memory includes: a U disk, a mobile hard disk, a magnetic disk or an optical disk, and various media that can store program codes.

[0094] The above is only exemplary embodiments of the present disclosure, which cannot limit the scope of the present disclosure. That is, any equivalent changes and modifications made in accordance with the teachings of the present disclosure are still within the scope of the present disclosure. Other embodiments of the present disclosure will be readily apparent to those skilled in the art upon considering the specification and practicing the true principles of the present disclosure.

[0095] The present application is intended to cover any variations, uses or adaptive changes of the present disclosure that follow the general principles of the present disclosure and include common knowledge or conventional technical means in the technical field not recorded in the present disclosure. The specification and examples are only considered as exemplary, and the scope and spirit of the present disclosure are defined by the claims.

Claims

1. A search engine based user profiling method, characterized in that, The method includes: Obtain user's target search behavior data; Based on the target search behavior data, determine the data corresponding to the user's static feature tags and dynamic feature tags; The first feature weight of the user is calculated based on the data corresponding to each of the dynamic feature tags; The first feature weights are corrected based on a preset time decay formula to obtain the second feature weights; A user profile of the user is generated based on each of the static feature labels and the second feature weights; The calculation of the user's first feature weight based on the data corresponding to each of the dynamic feature tags includes: From the dynamic feature tags, the search frequency tag, the search path tag, and the search click tag are selected; The weights of the search frequency tag, the search path tag, and the search click tag are calculated separately. Specifically, the weight of the search frequency tag is calculated as: Search frequency = Number of searches / Total number of searches. The number of searches refers to the number of times a user searches for a specific search target / keyword within a given period. The total number of searches refers to the sum of the number of searches for all search targets / keywords within a given period. The weight of the search path tag is calculated as: Number of clicks from manually entered searches / Total number of clicks. The number of clicks from manually entered searches refers to the number of times a user manually enters keywords to search. The total number of clicks refers to the sum of clicks from all search paths, including clicks from manually entered searches and clicks from various popular tags on the search interface. The weight of the search click tag is calculated as: Number of clicks on the search results page / Total number of clicks. The number of clicks on the search results page refers to the number of times a user clicks on a result related to the search keyword after performing a search. The total number of clicks refers to the sum of all clicks from the user on the search results page. The first feature weight is obtained by multiplying the weights of the search frequency tag, the search path tag, and the search click tag. The process of generating a user profile based on each of the static feature labels and the second feature weights includes: determining the user level according to the second feature weights and the correspondence between the weights and the level. Add the user level and the static feature tags to the user's user profile. 2.The search engine based user profiling method of claim 1, wherein, The acquisition of user target search behavior data includes: Obtain all user search behavior data; The search behavior data is cleaned to obtain cleaned search behavior data. Calculate the correlation between the data in the cleaned search behavior data; Data with relevance higher than the relevance threshold is filtered out to obtain target search behavior data. 3.The search engine based user profiling method of claim 1, wherein, Determining the user's static feature tags based on the target search behavior data includes: Obtain the keywords from the target search behavior data; The keywords are matched with a pre-set keyword library, and the keyword with the highest matching degree is selected as the target keyword. The target keyword is input into the trained user attribute prediction model to obtain predicted user attributes, and the user attributes are taken as the static feature labels. 4.The search engine based user profiling method of claim 1, wherein, The first feature weight is corrected based on a preset time decay formula to obtain a second feature weight, including: A current date and a previous date of the last search behavior data of the user are obtained; The current date, the previous date, and the first feature weight are substituted into the preset time decay formula to obtain a second feature weight. The preset time decay formula is Y=X*exp[-α(d1-d2)], where Y is the second feature weight, X is the first feature weight, α is an attenuation factor, d1 is the current date, and d2 is the previous date. 5.The search engine based user profiling method of claim 1, wherein, After the user portrait of the user is generated based on each static feature label and the second feature weight, the method further includes: Based on the user portrait, relevant information of interest to the user is determined; According to the user level, a corresponding information recommendation template is matched, and the relevant information is matched into the information recommendation template to obtain target recommendation information, and the information recommendation template includes at least one dimension of a related industry, a risk warning, a newly added enterprise, a newly added patent, and news updates. The target recommendation information is pushed to the user. 6.A search engine based user profiling system, characterized in that, A system for implementing the search engine-based user portrait construction method of claim 1, the system comprising: A search behavior data acquisition module (1) for acquiring target search behavior data of a user; A label determination module (2) for determining static feature labels and data corresponding to dynamic feature labels of the user according to the target search behavior data; A first feature weight calculation module (3) for calculating first feature weights of the user according to data corresponding to each dynamic feature label; A second feature weight correction module (4) for correcting the first feature weights based on a preset time decay formula to obtain second feature weights; A user portrait generation module (5) for generating a user portrait of the user based on each static feature label and the second feature weight.

7. A computer readable storage medium characterized by, The computer-readable storage medium stores a plurality of instructions, which are suitable for being loaded and executed by the processor to implement the method of any one of claims 1-5.

8. An electronic device, comprising: The electronic device includes a processor, a memory, and a transceiver, the memory is used to store instructions, the transceiver is used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to enable the electronic device to perform the method of any one of claims 1-5.

Images