WO2022262216A1 - Information recommendation method and device, and storage medium - Google Patents

Abstract

An information recommendation method. The method comprises: determining at least one first target tag corresponding to a sharing link operated by target user identification information in a current preset period (101); calculating a first number of user behaviors corresponding to each first target tag of the target user identification information in the current preset period (102); determining a corresponding time decay coefficient of the first target tag on the basis of each first number of user behaviors (103); performing user picture drawing on the basis of each first number of user behaviors and the corresponding time decay coefficient to obtain a picture drawing result of the target user identification information (104); and on the basis of the picture drawing result, determining a target recommendation link needing to be recommended to the target user identification information (105). Also disclosed are an information recommendation device and a storage medium.

Description

An information recommendation method, device and storage medium

Cross References to Related Applications

This application is based on a Chinese patent application with application number 202110660814.8 and a filing date of June 15, 2021, and claims the priority of this Chinese patent application. The entire content of this Chinese patent application is hereby incorporated by reference into this application.

technical field

The present application relates to the field of Internet application technology, and in particular to an information recommendation method, device and storage medium.

Background technique

With the rapid development of computer technology, more and more technologies are applied in the financial field, and the traditional financial industry is gradually transforming into financial technology (Fintech). However, due to the security and real-time requirements of the financial industry, more and more technical requirements high demands. With the rapid development of Internet technology, the Internet has had a major impact on people's life, study, entertainment, etc. After people see interesting content on the Internet, they can forward the interesting content and other operations. And some merchants can make statistics on the number of reposts and reading times of their published content, and then use a simple recommendation method to make targeted content recommendations for user behavior.

However, the currently commonly used recommendation methods are relatively simple and single, and there is a problem that the recommendation process cannot accurately count the user's preferences, resulting in the recommended content often not meeting the content expected by the user, resulting in a low accuracy rate of the recommended content.

Contents of the invention

In order to solve the above-mentioned technical problems, the embodiment of the present application expects to provide an information recommendation method, device and storage medium, which solves the problem that the current recommendation method is relatively simple and single, and the recommended content does not meet the user's expected content, and realizes a recommendation method , can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy.

The technical scheme of the present application is realized like this:

In the first aspect, an information recommendation method, the method includes:

Determine at least one first target tag corresponding to the sharing link operated by the target user identification information within the current preset period;

Counting the number of first user behaviors corresponding to each first target tag of the target user identification information within the current preset period; wherein, the first user behavior times are used to indicate that the target user identification information is shared with the target The estimated number of times the link performed an action;

Based on each of the first user behavior times, determine the corresponding time decay coefficient of the first target tag;

performing user portraits based on each of the first user behavior times and the corresponding time decay coefficients, to obtain a portrait result of the target user identification information;

Based on the portrait result, determine a target recommendation link to be recommended to the target user identification information.

In a second aspect, an information recommendation device includes: a memory, a processor, and a communication bus;

in:

The memory is used to store executable instructions;

The communication bus is used to realize the communication connection between the processor and the memory;

The processor is configured to execute the information recommendation program stored in the memory to implement the steps of the information recommendation method described in any one of the above.

According to a third aspect, a storage medium stores an information recommendation program, and when the information recommendation program is executed by a processor, the steps of the information recommendation method described in any one of the foregoing are implemented.

In the embodiment of the present application, after the information recommendation device determines at least one first target label corresponding to the sharing link operated by the target user identification information in the current preset period, it counts each first target of the target user identification information in the current preset period The number of first user behaviors corresponding to the label, and based on each first user behavior number, determine the corresponding time decay coefficient of each first target tag, and then perform User portrait, get the portrait result of the target user's identification information, and based on the portrait result, determine the target recommendation link that needs to be recommended to the target user's identification information. In this way, fully consider the situation that the user's preference and recommendation content will change over time Next, according to the number of first user behaviors and the time decay coefficient corresponding to at least one first target tag corresponding to the user in the current preset period, a user portrait is made for the user and the target recommendation link to be recommended is determined according to the user's portrait result , to solve the problem that the current recommendation method is relatively simple and single, which leads to the recommendation content not meeting the user's expectations, and implements a recommendation method that can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy.

Description of drawings

FIG. 1 is a schematic flow diagram of an information recommendation method provided in an embodiment of the present application;

FIG. 2 is a schematic flow chart of another information recommendation method provided by the embodiment of the present application;

FIG. 3 is a schematic flowchart of another information recommendation method provided by the embodiment of the present application;

FIG. 4 is a schematic flowchart of an information recommendation method provided by another embodiment of the present application;

FIG. 5 is a schematic structural diagram of an information recommendation device provided in an embodiment of the present application;

FIG. 6 is a schematic diagram of an application implementation process provided by an embodiment of the present application;

FIG. 7 is a schematic diagram of a user portrait radar provided by an embodiment of the present application;

FIG. 8 is a schematic structural diagram of another information recommendation device provided by an embodiment of the present application.

detailed description

In order to make the purpose, technical solutions and advantages of the application clearer, the application will be further described in detail below in conjunction with the accompanying drawings. All other embodiments obtained under the premise of creative labor belong to the scope of protection of this application.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of the application, and are not intended to limit the application.

An embodiment of the present application provides an information recommendation method. Referring to FIG. 1, the method is applied to an information recommendation device, and the method includes the following steps:

Step 101: Determine at least one first target tag corresponding to a sharing link operated by target user identification information within a current preset period.

In this embodiment of the application, the information recommendation device may be a management server that has a communication connection relationship with the user terminal device, and the user terminal device is a device that the user can view the content corresponding to the shared link, for example, it may be the user's smart mobile terminal or computer device, etc. . The sharing link can be a shopping link, or some article links, etc. The preset period can be an empirical value determined according to actual needs, through which statistical analysis can be performed to determine changes in user preferences, for example, it can be in units of hours, days, months, years, etc. The first target label is used to identify the category label to which the shared link belongs. For example, when the shared link is a shopping link, the first target label can be men’s clothing, women’s clothing, children’s clothing, shoes and hats, etc. When the shared link is an article link, the first target label A target label can be economy, entertainment, military, culture, etc.

Step 102: Count the number of first user behaviors corresponding to each first target tag in the target user identification information within the current preset period.

Wherein, the first number of user behaviors is used to indicate the estimated number of operations performed on the target sharing link by the target user identification information.

In this embodiment of the application, the target user identification information refers to the identification information used to uniquely identify the target user. For example, when the information recommendation device is a server corresponding to an instant messaging application, the target user identification information is the information recommendation device as the target The target account information assigned by the user. Since at least one first target label corresponding to the target user identification information in the current preset period has been determined, the number of first user behaviors corresponding to each first target label of the user corresponding to the target user identification information in the current period can be obtained by statistics .

Counting the number of first user behaviors corresponding to each first target label by the user corresponding to the target user identification information in the current cycle can be realized by the following steps: firstly, counting the user pair tags corresponding to the target user identification information in the current cycle is the first A target sharing link with a target label, and then count the different operation types and the number of operations of different operation types of the user corresponding to the target user identification information on the target sharing link in the current cycle, and finally, according to the quantification times corresponding to different operation types and The operation times of different operation types are quantitatively estimated to obtain the estimated times, and the estimated times are recorded as the first user behavior times.

Step 103: Determine the time decay coefficient of the corresponding first target tag based on each first user behavior times.

In the embodiment of the present application, the time decay coefficient of the first target tag is used to represent the change over time of the user's liking for the content corresponding to the first target tag over time.

Step 104: Perform user portrait based on each first user behavior number and corresponding time decay coefficient, and obtain a portrait result of target user identification information.

In the embodiment of the present application, user profile analysis is performed on each first user behavior frequency and corresponding time decay coefficient, so that a profile result including an analysis result corresponding to at least one first target tag can be obtained.

Step 105, based on the portrait result, determine the target recommendation link to be recommended to the target user's identification information.

In the embodiment of the present application, after determining the portrait result corresponding to the target user identification information, the target recommendation link that can be recommended to the target user identification information can be determined according to the portrait result corresponding to the target user identification information, because the target recommendation link is based on The portrait result of the target user identification information is determined, and the portrait result fully considers the time decay coefficient corresponding to the user's preference for a certain type of recommended content. Therefore, after the target recommendation link is recommended to the user corresponding to the target user identification information, The user accepts the content of the target recommended link to a higher degree, improves the recommendation efficiency, and ensures the user experience effect.

In the embodiment of the present application, after determining at least one first target label corresponding to the sharing link operated by the target user identification information in the current preset period, the corresponding first target label of the target user identification information in the current preset period is counted. The first user behavior times, and based on each first user behavior times, to determine the corresponding time decay coefficient of each first target tag, and then according to each first user behavior times and the corresponding time decay coefficient to perform user portrait , to obtain the portrait result of the target user's identification information, and based on the portrait result, determine the target recommendation link that needs to be recommended to the target user's identification information. In this way, in full consideration of the changes in user preferences over time, According to the number of first user behaviors and the time decay coefficient corresponding to at least one first target tag corresponding to the user in the current preset period, the user is profiled and the target recommendation link to be recommended is determined according to the user's profile result, to solve the problem To solve the problem that the current recommendation method is relatively simple and single, resulting in recommended content that does not meet user expectations, a recommendation method is implemented that can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy.

Based on the foregoing embodiments, the embodiments of the present application provide an information recommendation method. Referring to FIG. 2, the method is applied to an information recommendation device, and the method includes the following steps:

Step 201: Determine at least one first target tag corresponding to a sharing link operated by target user identification information within a current preset period.

In the embodiment of this application, the preset period is 7 days as an example for illustration. The current preset period is a complete time period closest to the current moment. For example, the current preset period is 7 days before the current moment. Information recommended The device counts at least one first target label corresponding to the sharing link operated by the target user identification information within 7 days before the current moment, assuming that the sharing links operated by the target user identification information are link 1, link 2 and link 3, corresponding to Yes, the tag corresponding to link 1 includes tag 1, the tag corresponding to link 2 includes tag 1 and tag 2, and the tag corresponding to link 3 includes tag 3. Therefore, it can be determined that the target user identification information corresponds to the third shared link. A target label is: label 1, label 2 and label 3.

Step 202: Determine at least one target sharing link operated by the target user identification information corresponding to each first target tag within the current preset period.

In this embodiment of the application, at least one target sharing link corresponding to label 1 is determined as link 1 and link 2, at least one target sharing link corresponding to label 2 is link 2, and at least one target sharing link corresponding to label 3 is link 3.

Step 203: Determine the operating parameters of the target user identification information for operating each target sharing link.

In the embodiment of the present application, the operation parameters for the user to operate on link 1, link 2 and link 3 are determined.

Step 204: Based on the operating parameters and preset weight coefficients of each target sharing link, determine the sub-action times for each target sharing link in the target user identification information, and obtain at least one sub-action number.

In the embodiment of the present application, the preset weight coefficients are weight values corresponding to different operating parameters, which are empirical values obtained from a large number of experiments. After determining the operating parameters of each target sharing link, if the operating parameters of a certain target sharing link include a parameter, the operating parameters of the target component link are multiplied by the preset weight coefficient corresponding to the operating parameter to obtain the target sharing link For the number of sub-behaviors, if the operating parameters of a certain target sharing link include at least two parameters, multiply each parameter by the corresponding preset parameter to obtain the product value, and then accumulate the product values corresponding to at least two parameters to obtain The number of times this goal shared the sub-action of the link. It should be noted that, in the case that the same target sharing link includes at least two tags, when calculating the number of sub-behaviors of different tags of the target sharing link, it is also necessary to multiply the corresponding weights of different tags.

Exemplarily, the operation parameters corresponding to link 1 include parameter 1 and parameter 2, the operation parameters corresponding to link 2 include parameter 3, and the operation parameters corresponding to link 3 include parameter 1, parameter 2 and parameter 3, wherein parameter 1 corresponds to The preset weight coefficient is coefficient 1, and the default weight coefficient corresponding to parameter 3 is coefficient 3. Therefore, it can be determined that the number of sub-behaviors corresponding to link 1 is: parameter 1*factor 1+parameter 2*factor 2, and link 3 corresponds to The number of sub-actions is: parameter 1*coefficient 1+parameter 2*coefficient 2+parameter 3*coefficient 3, since the labels corresponding to link 2 include label 1 and label 2, the corresponding weight of label 1 is 30%, and the weight of label 2 is 70%, therefore, for label 1 of link 2, the corresponding number of sub-actions is 30%*parameter 3*coefficient 3, and for label 2 of link 2, the number of corresponding sub-actions is 70%*parameter 3*coefficient 3.

Step 205. Determine the cumulative sum of at least one sub-behavior times to obtain the first user behavior times.

In other embodiments of the present application, at least one sub-behavior count corresponding to the target user identification information is accumulated and calculated to obtain the first user behavior count corresponding to the target user identification information. Exemplarily, the number of first user behaviors=the number of sub-behaviors corresponding to link 1+the number of sub-behaviors corresponding to link 2+the number of sub-behaviors corresponding to link 3=parameter 1*coefficient 1+parameter 2*coefficient 2+30%*parameter 3*coefficient 3+parameter 1*coefficient 1+parameter 2*coefficient 2+parameter 3*coefficient 3.

Step 206: Determine the time decay coefficient of the corresponding first target tag based on each first user behavior times.

In the embodiment of the present application, the number of times of each first user behavior is analyzed to determine whether the user's interest in the content of the corresponding first target tag has a decaying time decay coefficient within the current preset period. Exemplarily, the time decay coefficient corresponding to tag 1 is determined according to the first user behavior times corresponding to tag 1, the time decay coefficient corresponding to tag 2 is determined according to the first user behavior times corresponding to tag 2, and the time decay coefficient corresponding to tag 3 is determined according to the first user behavior times corresponding to tag 3 A number of user behaviors, determine the time decay coefficient corresponding to label 3.

Step 207 : Perform user portrait based on the number of first user behaviors and the corresponding time decay coefficient, and obtain the portrait result of the target user identification information.

In the embodiment of the present application, the user portrait processing is performed on each first user behavior number corresponding to the target user identification information and the corresponding time decay coefficient, and the portrait result corresponding to the target user identification information is obtained. Exemplarily, the first user behavior times corresponding to tag 1 and the time decay coefficient corresponding to tag 1 are analyzed to obtain the sub-portrait result of tag 1 corresponding to the target user identification information, and the first user behavior times corresponding to tag 2 and Analyze the time decay coefficient corresponding to tag 2 to determine the sub-portrait result of tag 2 corresponding to the target user identification information, analyze the first user behavior times corresponding to tag 3 and the time decay coefficient corresponding to tag 3, and determine the target user For the sub-portrait result of tag 3 corresponding to the identification information, determine the sub-portrait result of tag 1, the sub-portrait result of tag 2, and the sub-portrait result of tag 3 as the portrait result corresponding to the target user identification information. Wherein, the portrait result may be represented by a user portrait radar chart.

Step 208, based on the portrait result, determine the target recommendation link to be recommended to the target user's identification information.

In the embodiment of the present application, according to the portrait result corresponding to the target user identification information, select the link that matches the portrait result, obtain the target recommended link, and push the target recommended link to the user terminal device corresponding to the target user identification information, so that the user After the terminal device receives the target recommendation link, it can display the content corresponding to the target recommendation link, and the user corresponding to the target user identification information can click on the content corresponding to the target recommendation link to perform operations such as reading, forwarding, liking and/or commenting.

In this way, the operation parameters of the user's operation on each target sharing link are quantified to determine the number of sub-behaviors that the user operates on the target sharing link corresponding to each target label, which effectively improves the accuracy of user portraits for users. The accuracy rate ensures the accuracy of the target recommendation link recommended to the user.

Based on the aforementioned embodiments, in other embodiments of the present application, the operating parameters include at least one of the following parameters: reading operating parameters, reading duration parameters, forwarding operating parameters, liking operating parameters, and commenting operating parameters. Correspondingly, the operating parameters are reading When the duration parameter is used, the corresponding preset weight coefficient is the minimum value between the sum of the remainder between the reading duration parameter and the preset duration and 1 and the standard weight coefficient.

In this embodiment of the present application, in some cases, the forwarding operation parameter includes a sharing operation. When the operating parameter is the reading duration parameter, the process of determining the preset weight coefficient corresponding to the reading duration parameter is: calculate the sum of the reading duration parameter/remainder of the preset duration and 1, from the reading duration parameter/remainder of the preset duration+ 1 and the standard weight coefficient to determine the minimum value, so that the preset weight coefficients corresponding to different reading duration parameters can be effectively standardized and unified, and the reliability of the preset weight coefficients corresponding to different reading duration parameters can be improved. The standard weight coefficient can be an empirical value determined based on experience, or it can be determined based on actual experience by operation and maintenance developers. Correspondingly, the preset weight coefficients corresponding to other operating parameters except the reading duration parameter are all preset empirical values.

Based on the foregoing embodiments, in other embodiments of the present application, step 206 may be implemented by the following steps: counting the second corresponding to each first target tag of the target user identification information in the previous period adjacent to the current preset period User behavior times: based on the ratio between the first user behavior times corresponding to each first target tag and the corresponding second user behavior times, the time decay coefficient of the corresponding first target tag is obtained.

In this embodiment of the application, after calculating the ratio between the number of first user behaviors corresponding to each first target tag and the corresponding second user behavior times, the first user corresponding to each first target tag The ratio between the number of behaviors and the corresponding number of behaviors of the second user is processed by methods such as logarithmic calculation to obtain the time decay coefficient of the corresponding first target tag.

Based on the foregoing embodiments, in other embodiments of the present application, step 206 may also be implemented by steps 206a-206d:

Step 206a, counting the number of second user behaviors corresponding to each first target label of the target user identification information in the previous period adjacent to the current preset period.

In the embodiment of the present application, the implementation method of counting the number of second user behaviors corresponding to each first target label of the target user identification information in the previous period adjacent to the current preset period is the same as realizing the determination of the number of times in the current preset period The implementation method of the number of first user behaviors corresponding to each first target tag of the target user identification information is the same, and will not be described in detail here.

Exemplarily, according to the method of obtaining the first user behavior times of tag 1, the first user behavior times of tag 2, and the first user behavior times of tag 3 of the target user identification information through statistics, the statistically obtained In the previous period of the target user identification information, the number of second user behaviors corresponding to tag 1, the second user behavior times corresponding to tag 2, and the second user behavior times corresponding to tag 3.

Step 206b: Determine the logarithm of the ratio of the first user behavior times to the second user behavior times corresponding to each first target tag, and obtain a first value corresponding to each first target tag.

In the embodiment of the present application, when performing logarithmic calculation of the ratio of the first user behavior times to the second user behavior times corresponding to each first target tag, the natural logarithmic calculation method can be used to realize the logarithmic calculation process , that is to calculate the logarithmic value of the true number of the ratio of the first user behavior times to the second user behavior times corresponding to each first target tag with a constant e as the base. Exemplarily, the first numerical value for label 1=ln(the first user behavior number of label 1/the second user behavior number of label 1), the first numerical value for label 2=ln(the first user behavior number of label 2 times/second user behavior times of label 2), for the first numerical value of label 3=ln(first user behavior times of label 3/second behavior times of label 3).

Step 206c: Determine a first ratio of the first numerical value corresponding to each first target label to the cycle interval duration corresponding to the current preset cycle.

In the embodiment of the present application, the first ratio=the first numerical value corresponding to each first target tag/period interval duration. Exemplarily, since the period interval is 7 days, therefore, the first ratio of label 1 = the first value of label 1 / the period interval length = the first value of label 1 / 7, the first ratio of label 2 = label 2 The first value/period interval duration=the first value of tag 2/7, the first value/period interval duration of tag 3=the first value of tag 3/7.

Step 206d, determine the sum of the first ratio and 1 corresponding to each first target label, and obtain the time decay coefficient corresponding to each first target label.

In the embodiment of the present application, the time decay coefficient corresponding to each first target tag=the first ratio corresponding to each first target tag+1. Exemplarily, the time decay coefficient corresponding to label 1 of the target user identification information=the first ratio of label 1+1=the first numerical value of label 1/7+1=[ln(the number of first user behaviors of label 1/label 1 second user behavior times)]/7+1, the time decay coefficient corresponding to label 2 of the target user identification information=the first ratio of label 2+1=the first numerical value of label 2/7+1=[ln( The first user behavior times of label 2/the second user behavior times of label 2)]/7+1, the time decay coefficient corresponding to label 3 of target user identification information=the first ratio of label 3+1=the first ratio of label 3 A numerical value/7+1=[ln(the first user behavior times of label 3/the second user behavior times of label 3)]/7+1.

In this way, the time decay coefficient is obtained by analyzing the first user behavior times of each first target tag in the current preset cycle and the corresponding second user behavior times of the first target tag in the previous cycle, fully considering the Changes in preferences of users over time effectively improve the accuracy of subsequent user recommended content corresponding to the target user identification information.

Based on the aforementioned embodiments, in other embodiments of the present application, step 207 can be implemented by the following steps: determine the cumulative sum value of at least one first user behavior times to obtain the first sum value; determine the sum of each first user behavior times and The ratio of the first sum value to obtain the second ratio corresponding to each first target tag; count the number of users marked as each first target tag in the current preset period; a third ratio of the number of users of the target tag; determining a target logarithm for each third ratio; determining for each target logarithm, a second ratio corresponding to the same first target tag, and a time decay coefficient corresponding to the same first target tag product to obtain at least one first product; determine the portrait result corresponding to each target label as the corresponding first product.

Based on the foregoing embodiments, in other embodiments of the present application, step 207 may be implemented by steps 207a-207g:

Step 207a. Determine the cumulative sum of at least one first user behavior times to obtain a first sum.

In the embodiment of the present application, the cumulative sum value of at least one first user behavior number corresponding to the target user identification information is determined to obtain the first sum value. Exemplarily, the first sum value=the first user behavior times of label 1+the first user behavior times of label 2+the first user behavior times of label 3.

Step 207b. Determine the ratio of the number of times of each first user behavior to the first sum to obtain a second ratio corresponding to each first target label.

In the embodiment of the present application, the second ratio corresponding to each first target tag=the number of first user behaviors corresponding to each target tag/the first sum value. Exemplarily, the second ratio corresponding to label 1 = the number of first user behaviors corresponding to label 1 / the first sum value = the number of first user behaviors corresponding to label 1 / (the number of first user behaviors corresponding to label 1 + the number of first user behaviors of label 2 The number of times of the first user behavior + the number of times of the first user behavior of the label 3); the second ratio corresponding to the label 2 = the number of the first user behavior corresponding to the label 2 / the first sum value = the number of the first user behavior corresponding to the label 2 / ( The first user behavior times of label 1 + the first user behavior times of label 2 + the first user behavior times of label 3); the second ratio corresponding to label 3 = the first user behavior times corresponding to label 3 / the first sum value =the first user behavior times corresponding to label 3/(the first user behavior times of label 1+the first user behavior times of label 2+the first user behavior times of label 3).

Step 207c, counting the number of users marked as each first target label within the current preset period.

In the embodiment of the present application, the number of all users marked as each first target label corresponding to the target user identification information within the current preset period, wherein all users include the user corresponding to the target user identification information. Exemplarily, the number of users marked with label 1, the number of users marked with label 2 and the number of users marked with label 3 within the current preset period are counted.

Step 207d. Determine a third ratio of the number of actions of each first user to the number of users with the same first target label.

In the embodiment of the present application, the third ratio is the ratio of the number of first user behaviors of each first target tag to the number of users of the corresponding first target tag. Exemplarily, the third ratio corresponding to label 1 = the number of first user behaviors corresponding to label 1 / the number of users corresponding to label 1, the third ratio corresponding to label 2 = the number of first user behaviors corresponding to label 2 / the number of users corresponding to label 2 The third ratio corresponding to label 3 = the number of first user behaviors corresponding to label 3 / the number of users corresponding to label 3 .

Step 207e, determine the target logarithm of each third ratio.

In the embodiment of the present application, determining the target logarithm of each third ratio may be obtained by calculating the base 10, and the real number is the logarithm of each third ratio. Exemplarily, the target logarithm of label 1=lg the third ratio corresponding to label 1=lg (the number of first user behaviors corresponding to label 1/the number of users corresponding to label 1), the target logarithm of label 2=lg label 2 The corresponding third ratio=lg (the first user behavior times corresponding to label 2/the number of users corresponding to label 2), the target logarithm of label 3=lg the third ratio corresponding to label 3=lg (the first user behavior corresponding to label 3 Number of user behaviors/number of users corresponding to label 3).

Step 207f: Determine the product of each target logarithm, the second ratio corresponding to the same first target label, and the time decay coefficient corresponding to the same first target label, to obtain at least one first product.

In this embodiment of the present application, the first product corresponding to each first target label=the number of target pairs corresponding to each target label*the second ratio corresponding to the same first target label*the time decay coefficient corresponding to the same first target label. The quantity of at least one first product is the same as that of at least one first target label, and each first product corresponds to a first target label. Exemplarily, the first product corresponding to label 1 = the target logarithm corresponding to label 1 * the second ratio corresponding to label 1 * the time decay coefficient corresponding to label 2 .

Step 207g, perform user portrait based on at least one first product, and obtain a portrait result of target user identification information.

In this embodiment of the present application, a user portrait is performed on at least one first product corresponding to at least one first target tag to obtain a portrait result of target user identification information.

In this way, the operation parameters corresponding to each first target label corresponding to the target user identification information are specifically quantified, and the implementation process of user portraits that can be easily transplanted is realized, and the influence of the time decay coefficient is considered, and the portrait of the user portrait is improved. The accuracy of the result further ensures the accuracy of recommendation for the user corresponding to the target user identification information.

Based on the foregoing embodiments, in other embodiments of the present application, step 207g may be implemented by steps a11-a14:

Step a11. Determine the cumulative sum of at least one first product to obtain a second sum.

In this embodiment of the present application, at least one first product corresponding to at least one first target label is accumulated to obtain a second sum value. Exemplarily, the second sum value for target user identification information=the first product corresponding to label 1+the first product corresponding to label 2+the first product corresponding to label 3.

Step a12. Determine a fourth ratio of each first product to the second sum.

In the embodiment of the present application, the fourth ratio corresponding to each first target label=the first product/second sum corresponding to each first target label. Exemplarily, the fourth ratio corresponding to label 1 = the first product corresponding to label 1 / the second sum = the target logarithm corresponding to label 1 * the second ratio corresponding to label 1 * the time decay coefficient corresponding to label 2 / ( The first product corresponding to label 1 + the first product corresponding to label 2 + the first product corresponding to label 3).

Step a13. Determine the product of each fourth ratio and the preset magnification factor to obtain at least one second product.

In the embodiment of the present application, the preset magnification factor may be a factor required to magnify the fourth ratio according to actual needs, such as 1, 10 or 100. In this way, the fourth ratio of each first target tag is correspondingly amplified according to requirements, the number of digits after the decimal point can be reduced, and the user can quickly compare at least one second product, which is convenient for the user to compare at least one second product use to improve user experience.

Step a14, determining that the portrait result is at least one second product.

In the embodiment of the present application, the result of obtaining the portrait corresponding to the target user identification information is at least one second product. In this way, by obtaining the portrait results of each first target tag corresponding to the target user identification information and performing digital quantification, the target recommendation link corresponding to the target user identification information is effectively determined through digital analysis.

Based on the foregoing embodiments, in other embodiments of the present application, step 208 may be implemented by steps 208a-208d or steps 208a-208c and steps 208e-208k. Among them, if the first reference tag includes 1 tag, choose to execute steps 208a-208d, if the first reference tag includes at least two tags, choose to execute steps 208a-208c and steps 208e-208k:

Step 208a, from the at least one second product, determine a target product greater than or equal to the profile threshold.

In the embodiment of the present application, the image threshold is an empirical value obtained from a large number of experiments. Wherein, in at least one second product, there may be none of the second products greater than or equal to the portrait threshold, and there may be one or more second products greater than or equal to the portrait threshold, that is, the target product may be empty, or may include at least A second product greater than or equal to the portrait threshold.

Step 208b, from at least one first target label, determine a second target label corresponding to the target product.

In this embodiment of the present application, determining the target product greater than or equal to the portrait threshold from at least one second product, and determining the second target label corresponding to the target product from at least one first target label means determining that the target user identification information corresponds to content that users of may be more interested in during the current preset period.

Step 208c: Determine the first reference tag corresponding to the content of the link to be shared.

In this embodiment of the application, the first reference tag corresponding to the content of the link to be shared may be marked by the publisher of the link to be shared. In some application scenarios, the first reference tag corresponding to the content of the link to be shared The tag may also be determined after automatic analysis of the content of the link to be shared.

Step 208d, if the first reference tag includes 1 tag, and the first reference tag belongs to the second target tag, determine the target recommended link as the link to be shared.

In the embodiment of the present application, when the first reference tag only includes one tag, and the first reference identifier belongs to the second target tag, it can be determined that the target recommended link corresponding to the target user identification information includes the link to be shared.

Step 208e, if the first reference tag includes at least two tags, determine the tag weight of each tag in the first reference tag.

In this embodiment of the application, when the first reference tag includes at least two tags, that is, when the content of the link to be shared includes multiple tags, the tag weight of each tag in the first reference tag may be the release of the content of the link to be shared It may be obtained by pre-setting the weight for each tag, or it may be determined during the process of automatically tagging the content of the link to be shared.

Step 208f: Determine the proportional relationship between the tag weights of each tag included in the first reference tag.

In this embodiment of the application, assuming that the first reference tag includes reference tag 1, reference tag 2, and reference tag 3, the tag weights of each corresponding tag are tag weight 1, tag weight 2, and tag weight 3 in turn, and the corresponding The proportional relationship can be recorded as label weight 1:label weight 2:label weight 3.

In the embodiment of the present application, after the information recommendation device executes step 208f, it may choose to execute step 208g, or choose to execute steps 208h-208i, or choose to execute steps 208j-208k. Among them, if the proportional relationship is within the first proportional range, and at least one of the first reference tags belongs to the second target tag, choose to execute step 208g; if the proportional relationship is within the second proportional range, determine from the first reference tag When the first preset number of second reference tags with the largest tag weight ratio, you can choose to execute steps 208h to 208i; if the proportional relationship is within the second ratio range, determine from the first reference tags that the tag weight is greater than the weight threshold. When a second reference label is used, steps 208j-208k can be optionally performed:

Step 208g, if the proportional relationship is within the first proportional range, and at least one of the first reference tags belongs to the second target tag, determine the target recommended link as the link to be shared.

Wherein, the first ratio range is used to identify that the tag weight difference between tags included in the first reference tag is small.

In this embodiment of the present application, the first ratio range is used to indicate that the first reference tags include tag weights that are almost the same, for example, the first ratio range is a range close to 1:1:1. Exemplarily, assuming that tag weight 1:tag weight 2:tag weight 3=0.3:0.31:0.33, it belongs to the first proportion range, and when the reference tag 1 is the tag 1 corresponding to the target user identification information, the target recommended link can be determined for the link to be shared. If the ratio is within the first ratio range, but none of the tags in the first reference tags belong to the second target tag, then it is determined that the link to be shared is not pushed to the user corresponding to the target user identification information.

Step 208h, if the proportional relationship is within the second ratio range, determine the first preset number of second reference tags with the largest tag weight ratio from the first reference tags.

Wherein, the second ratio range is different from the first ratio range.

In the embodiment of the present application, the second ratio range is used to indicate a ratio range other than the first ratio range, that is, the difference between tag weights of each tag is relatively large. The preset number can be determined according to the number of tags included in the first reference tag, and the specific determination method can be determined according to the ratio of the number, for example, 30% of the number of tags included in the first reference tag is rounded up, For example, when the number of tags included in the first reference tag is 2, the corresponding preset number is 1; when the number of tags included in the first reference tag is 3, the corresponding preset number is 1, and the tag included in the first reference tag When the quantity of is 4, the corresponding preset quantity is 2.

Step 208i, if at least one of the preset number of second reference tags belongs to the second target tag, determine the target recommended link as the link to be shared.

Step 208j, if the proportional relationship is within the second proportional range, determine at least one second reference tag whose tag weight is greater than the weight threshold from the first reference tags.

Wherein, the second ratio range is different from the first ratio range.

In the embodiment of the present application, the weight threshold may be set by the user himself, or may be an empirical value obtained from a large number of experiments used to indicate that beyond the weight threshold, the user has a higher probability of liking the content of the tag.

Step 208k, if at least one of the at least one second reference tag belongs to the second target tag, determine the target recommended link as the link to be shared.

In this way, the target recommendation link is determined by targeting, that is, whether to share to the user corresponding to the target user identification information through a tag in the link to be shared, or to select a large-scale target from at least two tags in the link to be shared The user identification information realizes a wider range of recommendations, but maintains that the recommended content is relatively in line with the user's preferences, effectively ensuring the user's experience.

Based on the foregoing embodiments, in other embodiments of the present application, as shown in FIG. 3 , before the information recommendation device performs step 201, it may also choose to perform steps 209-213:

Step 209, if it is detected that the target user identification information is for the target operation of the currently displayed link content, and the target operation is a permitted operation, determine the target parameter corresponding to the target operation.

Among them, the target parameter belongs to the operation parameter.

In this embodiment of the application, when it is detected that the target user identification information is operating on the target currently displaying the link content, the corresponding allowed operation may be that the target user identification information is in the whitelist user database, or the target user identification information is not in the blacklist user information library, or, when the target operation is a read operation, the current number of reads is lower than the upper limit of allowed reading, or when the target operation is a forwarding operation, the current number of forwardings is lower than the upper limit of allowed forwarding.

Step 210: Generate operation identification information for identifying the target operation, and store the operation identification information in a target storage area corresponding to the target user identification information.

In the embodiment of the present application, the target storage area may be a database for storing data, such as a distributed file storage database, or a data warehouse tool (Hive) library. The operation is information is used to uniquely identify the operation of the user corresponding to the target user identification information on the currently displayed link content, and store the operation identification information corresponding to the target operation in the target storage area to prevent it from being modified by criminals.

Step 211, if it is determined that the reference user identification information is obtained, store the reference user identification information in a target storage area corresponding to the target user identification information.

Wherein, the reference user identification information is the identification information of the user who shares the currently displayed link content with the target user identification information.

Step 212, identifying the parent node of the target user ID information as the reference user ID information.

Step 213, if it is not determined that the reference user ID information is obtained, the parent node for identifying the target user ID information is empty.

In the embodiment of the present application, according to the stored information of the parent node, it may be determined to perform a link forwarding process for the currently displayed link content.

In this way, each operation is recorded in the target storage area, and it is possible to accurately know who clicked to read or repost based on the operation records, narrowing down the scope of potential users for subsequent precise 1-to-1 promotion and marketing, and can be configured homomorphically at any time The number of link clicks and reading or forwarding permissions can achieve the effect of blocking users, and the publisher can see all the operating parameters of the sharing link, and conduct promotion analysis based on the operating parameters, watch the effect of the sharing link, and replace the product in time if the effect is not good Copywriting or product category, etc.

Based on the aforementioned embodiments, in other embodiments of the present application, as shown in FIG. 4 , after the information recommendation device executes step 211, it is also used to execute step 214 and/or step 215; wherein, for the currently displayed link content, the user only executes When the reading operation is performed, step 214 is selected to be performed; when the user only performs the forwarding operation, step 215 is selected to be performed; when the user performs the forwarding operation while the reading operation is performed, step 214 and step 215 are selected to be performed:

Step 214, if the target operation includes a reading operation parameter, determine the reading count of the currently displayed link content as the current reading count of the currently displayed link content plus 1.

In the embodiment of the present application, the reading times are updated to effectively ensure the reliability of the reading authority.

Step 215: If the target operation includes forwarding operation parameters, determine the forwarding times of the currently displayed link content as the current forwarding times of the currently displayed link content plus 1.

In the embodiment of the present application, the forwarding is updated here to effectively ensure the reliability of the forwarding weight.

Based on the foregoing embodiments, this embodiment of the present application provides a schematic structural diagram of an information recommendation device. Referring to FIG. 5 , the information recommendation device includes: a trajectory recording module 31, a user portrait module 32, and an intelligent recommendation module 33; wherein:

Track record module 31 implementation process can refer to shown in Figure 6, comprise the following steps:

Step 41. The information recommendation device detects that the publisher publishes a sharing link.

Among them, after the publisher publishes the sharing link, the information recommendation device will share the sharing track identification information for the sharing link, the sharing link identification information, the user identification information of the sharing user, the parent node information, and the operation parameter type (for example, use 1 to indicate the sharing operation , 2 means read operation, 3 means forward operation), and create information such as the operation time of each operation and store it in the database. Among them: there is no parent node for the initial sharing. After other users click and forward the sharing link, relevant information of the clicked user will be obtained through webpage authorization, except for the user's user identification information (Identity Document, ID), user nickname, gender, and province. In addition to the combination information of one or more of information such as , country, avatar picture, etc., the information shown in Table 1 may also be included.

It should be noted that when storing and processing in the database, the parent node ID of the currently clicked user is the user ID shared with the currently clicked user. If the currently clicked user shares or reads the shared link, the information recommendation device will use the current Clicking on the user's operation generates a corresponding operation ID. For example, when the current clicking user shares the sharing link, a new sharing track ID is generated. The track ID of each operation track point is unique.

Table 1

Step 42, the information recommendation device detects that user A clicks on the sharing link.

Wherein, the information recommendation device detects that user A clicks on the sharing link, which may be obtained by the information recommendation device monitoring the communication operation of user A's user terminal device.

Step 43 , the information recommendation device judges whether user A has reading permission, if user A has reading permission, execute step 44 , if user A does not have reading permission, execute step 411 .

Wherein, the information recommending device can determine whether user A has reading permission by judging whether user A's user ID is in the blacklist. Or it may also be determined according to the relationship between the current number of reading times and the upper limit of allowed reading.

Step 44, the information recommendation device judges whether there are remaining clicks on the sharing link, if there are remaining clicks, perform step 45, and if there are no remaining clicks, perform step 411.

Step 45: The information recommendation device responds to the user A's click operation, and allows the user terminal device of the user A to access the details page corresponding to the sharing link.

Step 46, the information recommendation device detects that user A clicks the share button.

Step 47, the information recommendation device judges whether user A has sharing authority, if user A has sharing authority, execute step 48, if user A does not have sharing authority, execute step 49.

Wherein, the information recommendation device may determine whether user A has the sharing authority through the sharing blacklist, or may also determine according to the relationship between the current number of sharing times and the allowed sharing upper limit.

Step 48: The information recommendation device responds to the user A's operation of clicking the share button, enabling user A to successfully share the content corresponding to the link for sharing.

Step 49, the information recommending device sends the indication information without sharing permission to the user terminal device.

In step 410, when the information recommendation device detects that user B clicks on the sharing link shared by user A, step 43 is repeated.

Step 411, the information recommendation device guides the user terminal device to display a blank page or a sorry page.

Wherein, when the information recommendation device performs authority control, it can be realized through the information in Table 2. The user portrait module 32, assuming that it is used to make a user portrait for user A, can go through the processes corresponding to steps 201-208, steps 206a-206d, steps 207a-207g, steps a11-a14, and steps 208a-208i. Exemplarily, the preset weight coefficient corresponding to the operation parameter can refer to Table 3, wherein, the preset weight coefficient corresponding to the reading duration parameter, T represents the reading duration parameter, the unit is second (s), and the preset duration is 20s, min(T/20+1, 3) means to determine a minimum value between T/20+1 and the standard weight coefficient 3.

Table 2

Share Track ID	Fine-grained configuration, only for a single sharing limit
Resource ID	Coarse-grained configuration, limit the number of times and blacklist for a certain resource
read limit	The maximum number of readings for this share, after reaching it, you can no longer read it
forwarding limit	The maximum number of reposts for this share, after reaching the maximum number of reposts, it cannot be reposted
blacklist	pull blacklist

table 3

Exemplarily, assuming that statistics are obtained within the current preset period, the user data corresponding to the entire database can be shown in Table 4, wherein the first column represents the user identification information of the user, the first row represents the label identification information, and the second to The four columns represent the corresponding first user behavior times under different labels, which are calculated according to the specific implementation process of steps 204 to 204 in the foregoing embodiment, and will not be described in detail here.

Table 4

the	label T1	label T2	label T3	label T4	...
User P1	5	2	1	0	...
User P2	10	3	0	0	...
User P3	0	0	5	4	...
User P4	0	20	0	0	...
User P5	0	15	2	4	...

Correspondingly, taking user P1 as an example, the second ratio TF for tag 1 (user P1, tag T1)=5/(5+2+1+0)=0.625; the corresponding target object IDF (user P1, tag T1 )=lg(5/(2+1))=0.222; use big data to analyze the user behavior data of the previous day every morning, and take the data of the latest week as the time interval, where the time interval can be fine-tuned according to the business scenario or randomly, assuming The number of times user P1 acted on tag T1 in the first 7 days was 40. On the second day, the number of times of user P1’s behavior on tag T1 in the first 7 days was 50 through big data statistics. At this time, the corresponding time decay coefficient λ=1+ln (50/40)/7=1+1.223/7=1.17; finally, it can be determined that the current real weight of tag T1 for user P1 is the first product θ ₁ =TF×IDF×λ=0.625×0.222×1.17=0.163 .

For user P1, four first products will be obtained, namely θ ₁ , θ ₂ , θ ₃ , and θ ₄ label weights. Assuming that the preset method multiple is 10, the corresponding second product for each label of user P1 The product in turn is

Exemplarily, assuming that user P1 includes 5 tags, α ₁ = 0.4 corresponding to tag 1 of user P1, α ₂ = 3.1 corresponding to tag 2 of user P1, and α ₃ =1.8, α ₄ =2.9 corresponding to label 4 of user P1, and α ₅ =1.8 corresponding to label 5 of user P1. The user radar image as shown in FIG. 7 can be used to represent the 5 second products of user P1.

For user P1, four first products will be obtained, namely θ ₁ , θ ₂ , θ ₃ , and θ ₄ label weights. Assuming that the preset method multiple is 10, the corresponding second product for each label of user P1 The product in turn is

Exemplarily, assuming that user P1 includes 5 tags, α ₁ = 0.4 corresponding to tag 1 of user P1, α ₂ = 3.1 corresponding to tag 2 of user P1, and α ₃ =1.8, α ₄ =2.9 corresponding to label 4 of user P1, and α ₅ =1.8 corresponding to label 5 of user P1. The user radar image as shown in FIG. 7 can be used to represent the 5 second products of user P1.

The intelligent recommendation module 33 is used to implement the specific implementation process corresponding to step 208, which will not be described in detail here.

It should be noted that, for descriptions of the same steps and content in this embodiment as in other embodiments, reference may be made to the descriptions in other embodiments, and details are not repeated here.

In the embodiment of the present application, after determining at least one first target label corresponding to the sharing link operated by the target user identification information in the current preset period, the corresponding first target label of the target user identification information in the current preset period is counted. The first user behavior times, and based on each first user behavior times, to determine the corresponding time decay coefficient of each first target tag, and then according to each first user behavior times and the corresponding time decay coefficient to perform user portrait , to obtain the portrait result of the target user's identification information, and based on the portrait result, determine the target recommendation link that needs to be recommended to the target user's identification information. In this way, in full consideration of the changes in user preferences over time, According to the number of first user behaviors and the time decay coefficient corresponding to at least one first target tag corresponding to the user in the current preset period, the user is profiled and the target recommendation link to be recommended is determined according to the user's profile result, to solve the problem To solve the problem that the current recommendation method is relatively simple and single, resulting in recommended content that does not meet user expectations, a recommendation method is implemented that can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy.

Based on the foregoing embodiments, an embodiment of the present application provides an information recommendation device. Referring to FIG. 8, the information recommendation device 5 may include: a processor 51, a memory 52, and a communication bus 53, wherein:

Memory 52, used to store executable instructions;

Communication bus 53, for realizing the communication connection between processor 51 and memory 52;

The processor 51 is configured to execute the information recommendation program stored in the memory 52, so as to realize the following steps:

Determine at least one first target tag corresponding to the sharing link operated by the target user identification information within the current preset period;

Counting the number of first user behaviors corresponding to each first target tag of the target user identification information within the current preset period; wherein, the first user behavior times are used to represent the estimated number of times the target user identification information operates on the target sharing link;

Determine the time decay coefficient of the corresponding first target tag based on each first number of times of user behavior;

performing user portraits based on the number of times of each first user behavior and the corresponding time decay coefficient, and obtaining a portrait result of target user identification information;

Based on the portrait result, determine the target recommendation link that needs to be recommended to the target user's identification information.

In other embodiments of the present application, when the processor 51 executes the step of counting the number of first user behaviors corresponding to each first target tag of the target user identification information within the current preset period, it may be implemented through the following steps:

Determine at least one target sharing link operated by the target user identification information corresponding to each first target tag within the current preset period;

Determine the operating parameters for the target user identification information to operate on each target sharing link;

Based on the operating parameters and preset weight coefficients of each target sharing link, determine the number of sub-behaviors of the target user identification information for each target sharing link, and obtain at least one number of sub-behaviors;

Determine the cumulative sum of at least one sub-behavior times to obtain the first user behavior times.

In other embodiments of the present application, the operation parameters include at least one of the following parameters: reading operation parameters, reading duration parameters, forwarding operation parameters, like operation parameters, and comment operation parameters. Correspondingly, when the operation parameters are reading duration parameters, the corresponding The preset weight coefficient of is the minimum value between the sum of the remainder between the reading duration parameter and the preset duration and 1 and the standard weight coefficient.

In other embodiments of the present application, when the processor 51 executes the step of determining the time decay coefficient of the corresponding first target tag based on the number of times of each first user behavior, it may be implemented through the following steps:

Counting the number of second user behaviors corresponding to each first target label of the target user identification information in the previous period adjacent to the current preset period;

Determine the logarithm of the ratio of the first user behavior times to the second user behavior times corresponding to each first target label, and obtain the first value corresponding to each first target label;

Determine the first ratio of the first numerical value corresponding to each first target label to the cycle interval duration corresponding to the current preset cycle;

The sum of the first ratio and 1 corresponding to each first target label is determined to obtain a time decay coefficient corresponding to each first target label.

In other embodiments of the present application, when the processor 51 executes the step of performing user portraits based on the number of first user behaviors and corresponding time decay coefficients, and obtains the portrait results of the target user identification information, it may be implemented through the following steps:

Determining the cumulative sum of at least one number of first user behaviors to obtain a first sum;

Determining the ratio of the number of times of each first user behavior to the first sum value to obtain a second ratio corresponding to each first target label;

Count the number of users marked as each first target label within the current preset period;

Determine the third ratio of the number of times of each first user behavior to the number of users with the same first target label;

determining a target logarithm for each third ratio;

determining the product of each target logarithm, the second ratio corresponding to the same first target label, and the time decay coefficient corresponding to the same first target label to obtain at least one first product;

Perform user portrait based on at least one first product to obtain a portrait result of target user identification information.

In other embodiments of the present application, when the processor 51 executes the step of performing user portrait based on at least one first product, and obtains the portrait result of the target user identification information, it may be implemented through the following steps:

determining a cumulative sum of at least one first product to obtain a second sum;

determining a fourth ratio of each first product to the second sum;

determining a product of each fourth ratio and a predetermined magnification to obtain at least one second product;

Determining the profiling result is at least one second product.

In other embodiments of the present application, when the processor 51 executes the step of determining the target recommendation link recommended to the target user identification information based on the portrait result, it may be implemented through the following steps:

from the at least one second product, determining a target product greater than or equal to the profile threshold;

From at least one first target label, determine a second target label corresponding to the target product;

determining the first reference tag corresponding to the content of the link to be shared;

If the first reference tag includes 1 tag, and the first reference tag belongs to the second target tag, determine the target recommended link as the link to be shared.

In other embodiments of the present application, after the processor 51 executes the step of determining the first reference tag corresponding to the content of the link to be shared, it is further configured to execute the following steps:

If the first reference label includes at least two labels, determine the label weight of each label in the first reference label;

Determining the proportional relationship between the label weights of each label included in the first reference label;

If the ratio is within the first ratio range, and at least one tag in the first reference tag belongs to the second target tag, determine that the target recommended link is the link to be shared; wherein, the first ratio range is used to identify the first reference tag included The label weight difference before the label is small;

If the proportional relationship is within the second ratio range, determine at least one second reference tag whose tag weight has the largest ratio and whose tag weight is greater than the weight threshold from the first reference tag; wherein, the second ratio range is different from the first ratio range ;

If at least one of the at least one second reference tag belongs to the second target tag, the target recommended link is determined as the link to be shared.

In other embodiments of the present application, before the processor 51 executes the step of determining at least one first target tag corresponding to the sharing link operated by the target user identification information within the current preset period, it is further configured to perform the following steps:

If it is detected that the target user identification information is a target operation on the currently displayed link content, and the target operation is a permitted operation, determine the target parameters corresponding to the target operation; wherein, the target parameters belong to the operation parameters;

Generate operation identification information for identifying the target operation, and store the operation identification information in the target storage area corresponding to the target user identification information;

If it is determined that the reference user identification information is obtained, store the reference user identification information to the target storage area corresponding to the target user identification information; wherein, the reference user identification information is the identification information of the user who shares the currently displayed link content with the target user identification information;

The parent node identifying the target user identification information is the reference user identification information;

If the reference user identification information is not determined to be obtained, the parent node for identifying the target user identification information is empty.

In other embodiments of the present application, if the processor 51 executes the step of determining that the reference user identification information is obtained, after storing the reference user identification information in the target storage area corresponding to the target user identification information, it is also used to perform the following steps:

If the target operation includes reading operation parameters, determine the reading times of the currently displayed link content as the current reading times of the currently displayed link content plus 1;

And/or, if the target operation includes a forwarding operation parameter, it is determined that the number of forwarding times of the currently displayed link content is the current number of forwarding times of the currently displayed link content plus 1.

It should be noted that one or more programs in the embodiments of the present application can be explained by the steps of one or more processors, and the implementation process of the method provided by the embodiments corresponding to Figures 1 to 4 can be referred to, and details will not be repeated here. . Wherein, the information recommendation device shown in FIG. 5 and the information recommendation device 5 are the same device.

In the embodiment of the present application, after the information recommendation device determines at least one first target label corresponding to the sharing link operated by the target user identification information in the current preset period, it counts each first target of the target user identification information in the current preset period The number of first user behaviors corresponding to the label, and based on each first user behavior number, determine the corresponding time decay coefficient of each first target tag, and then perform User portrait, get the portrait result of the target user's identification information, and based on the portrait result, determine the target recommendation link that needs to be recommended to the target user's identification information. In this way, fully consider the situation that the user's preference and recommendation content will change over time Next, according to the number of first user behaviors and the time decay coefficient corresponding to at least one first target tag corresponding to the user in the current preset period, a user portrait is made for the user and the target recommendation link to be recommended is determined according to the user's portrait result , to solve the problem that the current recommendation method is relatively simple and single, resulting in the recommended content not meeting the user's expectations, and implement a recommendation method that can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy.

Based on the foregoing embodiments, the embodiments of the present application provide a computer-readable storage medium, referred to as a storage medium for short, where one or more programs are stored in the computer-readable storage medium, and the one or more programs can be used by one or more The processor executes to implement the implementation process of the information recommendation method provided in the embodiments corresponding to FIGS. 1 to 4 , which will not be repeated here.

The above are only examples of the present application, and are not intended to limit the scope of protection of the present application. Any modifications, equivalent replacements and improvements made within the spirit and scope of the present application are included in the protection scope of the present application.

Industrial Applicability

An embodiment of the present application provides an information recommendation method, device, and storage medium. The method includes: determining at least one first target tag corresponding to a sharing link operated by target user identification information within the current preset period; counting the current preset The number of first user behaviors corresponding to each first target tag of the target user identification information in the period; based on each of the first user behavior times, determine the corresponding time decay coefficient of the first target tag; Performing user portraits on the first user behavior times and the corresponding time decay coefficients to obtain the portrait results of the target user identification information; based on the portrait results, determine the target recommendations that need to be recommended to the target user identification information link, which solves the problem that the current recommendation method is relatively simple and single, resulting in recommended content that does not meet the user's expectations. It implements a recommendation method that can accurately profile user behavior, accurately recommend content for target users, and improve recommendation accuracy. .

Claims (12)

1. A method for recommending information, the method comprising:

   Determine at least one first target tag corresponding to the sharing link operated by the target user identification information within the current preset period;

   Counting the number of first user behaviors corresponding to each first target tag of the target user identification information within the current preset period;

   Based on each of the first user behavior times, determine the corresponding time decay coefficient of the first target tag;

   performing user portraits based on each of the first user behavior times and the corresponding time decay coefficients, to obtain a portrait result of the target user identification information;

   Based on the portrait result, determine a target recommendation link to be recommended to the target user identification information.
2. The method according to claim 1, wherein the counting the number of first user behaviors corresponding to each first target tag of the target user identification information within the current preset period includes:

   determining at least one target sharing link operated by the target user identification information corresponding to each of the first target tags within the current preset period;

   Determining the operating parameters for the target user identification information to operate on each of the target sharing links;

   Based on the operating parameters and preset weight coefficients of each of the target sharing links, determine the number of sub-behaviors of the target user identification information for each of the target sharing links to obtain at least one number of sub-behaviors;

   Determining the cumulative sum of the at least one sub-behavior times to obtain the first user behavior times.
3. The method according to claim 2, wherein the operation parameters include at least one of the following parameters: reading operation parameters, reading duration parameters, forwarding operation parameters, like operation parameters and comment operation parameters, correspondingly, the operation parameters are reading When the duration parameter is used, the corresponding preset weight coefficient is the minimum value between the sum of the remainder between the reading duration parameter and the preset duration and 1 and the standard weight coefficient.
4. The method according to claim 1, wherein said determining the time decay coefficient of the corresponding first target tag based on each of said first user behavior times includes:

   Counting the number of second user behaviors corresponding to each of the first target tags of the target user identification information in the previous cycle adjacent to the current preset cycle;

   Determining the logarithm of the ratio of the number of times of the first user behavior corresponding to each of the first target tags to the number of times of the second user behavior to obtain a first value corresponding to each of the first target tags;

   determining a first ratio of the first numerical value corresponding to each of the first target tags to the cycle interval duration corresponding to the current preset cycle;

   Determining the sum of the first ratio and 1 corresponding to each of the first target tags to obtain the time decay coefficient corresponding to each of the first target tags.
5. The method according to any one of claims 1 to 4, wherein the user portrait is performed based on each of the first user behavior times and the corresponding time decay coefficient to obtain the portrait result of the target user identification information, include:

   Determining at least one cumulative sum value of the number of times of the first user behavior to obtain a first sum value;

   Determining the ratio of each of the first user behavior times to the first sum to obtain a second ratio corresponding to each of the first target tags;

   counting the number of users marked as each of the first target tags within the current preset period;

   determining a third ratio of the number of times of each first user behavior to the number of users with the same first target label;

   determining a target logarithm for each of said third ratios;

   determining a product of each of the target logarithms, the second ratio corresponding to the same first target label, and the time decay coefficient corresponding to the same first target label to obtain at least one first product;

   Perform user portrait based on at least one of the first products to obtain a portrait result of the target user identification information.
6. The method according to claim 5, wherein the user portrait based on at least one of the first products to obtain the portrait result of the target user identification information includes:

   determining a cumulative sum of said at least one of said first products to obtain a second sum;

   determining a fourth ratio of each of said first products to said second sum;

   determining a product of each of said fourth ratios and a predetermined magnification to obtain at least one second product;

   Determining the portrait result as the at least one second product.
7. The method according to claim 6, wherein said determining a target recommendation link recommended to the target user identification information based on the portrait result comprises:

   from said at least one second product, determining a target product greater than or equal to a profile threshold;

   From the at least one first target label, determine a second target label corresponding to the target product;

   determining the first reference tag corresponding to the content of the link to be shared;

   If the first reference tag includes 1 tag, and the first reference tag belongs to the second target tag, determine that the target recommended link is the link to be shared.
8. The method according to claim 7, wherein, after determining the first reference tag corresponding to the content of the link to be shared, the method further comprises:

   If the first reference tag includes at least two tags, determining the tag weight of each tag in the first reference tag;

   determining the proportional relationship between the label weights of each label included in the first reference label;

   If the proportional relationship is within the first proportional range, and at least one of the first reference tags belongs to the second target tag, determine the target recommended link as the link to be shared; wherein, the second A ratio range is used to identify that the label weight difference before the label included in the first reference label is small;

   If the proportional relationship is within the second ratio range, determining the first preset number of second reference tags with the largest tag weight ratio from the first reference tags;

   If at least one of the preset number of second reference tags belongs to the second target tag, determining the target recommended link as the link to be shared;

   Alternatively, if the proportional relationship is within a second proportional range, at least one second reference tag whose tag weight is greater than a weight threshold is determined from the first reference tags; wherein, the second proportional range is the same as the first proportional different range;

   If at least one tag in the at least one second reference tag belongs to the second target tag, determine the target recommended link as the link to be shared.
9. The method according to claim 3 or 4, wherein, before determining at least one first target label corresponding to the sharing link operated by the target user identification information within the current preset period, the method further includes:

   If it is detected that the target user identification information performs a target operation on the currently displayed link content, and the target operation is a permitted operation, determine the target parameter corresponding to the target operation; wherein the target parameter belongs to the operation parameter;

   generating operation identification information for identifying the target operation, and storing the operation identification information in a target storage area corresponding to the target user identification information;

   If it is determined that the reference user identification information is obtained, store the reference user identification information in the target storage area corresponding to the target user identification information; wherein, the reference user identification information is to share the currently displayed link content with the target user identification Identification information of the user of the information;

   identifying the parent node of the target user identification information as the reference user identification information;

   If the reference user identification information is not determined to be obtained, the parent node identifying the target user identification information is empty.
10. The method according to claim 9, wherein, if it is determined that the reference user identification information is obtained, after storing the reference user identification information in the target storage area corresponding to the target user identification information, the method further comprises:

    If the target operation includes the reading operation parameter, determine that the number of readings of the currently displayed link content is the current number of readings of the currently displayed link content plus 1;

    And/or, if the target operation includes a forwarding operation parameter, determine that the number of forwarding times of the currently displayed link content is the current number of forwarding times of the currently displayed link content plus 1.
11. An information recommendation device, the device comprising: a memory, a processor, and a communication bus; wherein:

    The memory is used to store executable instructions;

    The communication bus is used to realize the communication connection between the processor and the memory;

    The processor is configured to execute the information recommendation program stored in the memory to implement the steps of the information recommendation method according to any one of claims 1-10.
12. A storage medium, an information recommendation program is stored on the storage medium, and when the information recommendation program is executed by a processor, the steps of the information recommendation method according to any one of claims 1 to 10 are realized.

Images