Application promotion service intelligent optimization system based on data analysis

Abstract

The invention relates to the technical field of data analysis, in particular to an application promotion service intelligent optimization system based on data analysis, which comprises a demand acquisition module, a demand matching module, an interaction acquisition module, an interaction analysis module and a comprehensive analysis module. The interaction judgment unit judges whether a user is idle or busy, the interaction collection unit collects interaction information, the interaction analysis module carries out data analysis on the collected interaction information, an interaction feature data set is generated, and a feature evaluation value is obtained according to a standard feature model. A personal interaction training model is constructed through the feature evaluation value, a first feature interval, a second feature interval and a third feature interval are generated, comprehensive judgment is performed through a comprehensive analysis module in combination with an interaction state and different feature intervals, a judgment result is obtained, popularization is performed according to the judgment result, a server-to-end popularization mode is optimized, and the popularization efficiency is improved. And the intelligent popularization effect is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to an intelligent optimization system for application promotion services based on data analysis. Background Art

[0002] App promotion is an advertising technology based on big data analysis of apps. It obtains app data information and provides promotion services to users based on their needs.

[0003] In the prior art, such as in the patent document entitled A Technology Promotion System and Method Based on the Internet with application number CN202210436234.5, the corresponding attribute factors and exposure factors are obtained by combining and comprehensively evaluating various data on the own attributes and real-time promotion exposure of different technologies, and the inferred valuations are obtained by combining the attribute factors and the exposure factors. Based on the inferred valuations, different types of technologies can be comprehensively evaluated from different aspects, and matching analysis of the inferred valuations can realize the classification and sorting of technologies of the same type, so as to facilitate the subsequent dynamic promotion of different technologies in an adaptive manner. This invention can solve the problem of poor overall effect of data-level processing and allocation for the promotion of different technologies in the existing solutions, but the promotion function described in the above patent document cannot generate interactive information with users, and it is difficult to promote according to personal habits. In terms of server-to-end, it is not possible to perform intelligent promotion based on personal data. Summary of the invention

[0004] The purpose of the present invention is to solve the problems in the background technology and to propose an application promotion service intelligent optimization system based on data analysis.

[0005] The present invention adopts the following technical scheme: an intelligent optimization system for application promotion services based on data analysis, including a demand acquisition module, a demand matching module, an interactive collection module, an interactive analysis module and a comprehensive analysis module;

[0006] The demand acquisition module is used to obtain the user's demand information and transmit it to the application data analysis module. The demand information includes search keywords and clicked tag types; the search keywords are obtained through the search box;

[0007] The demand matching module is used to obtain application information in the server, including application ratings and application loads, match the application information with the demand information, obtain information matching results, generate a sequence to be promoted based on the information matching results, and upload it to the comprehensive analysis module;

[0008] An interaction collection module, including an interaction state judgment unit and an information collection unit;

[0009] The interaction status determination unit is used to determine whether the user's interaction status is idle or busy;

[0010] The information collection unit is used to obtain the user's interactive information and transmit it to the comprehensive analysis module;

[0011] The interaction analysis module obtains the interaction status obtained in the interaction acquisition module, and performs interaction feature processing on the interaction information to obtain an interaction feature data set, inputs the interaction feature data set into the standard feature model for feature evaluation, obtains a feature evaluation result, generates a personal interaction training model according to the feature evaluation result and the interaction status, defines a feature interval through the personal interaction training model, and transmits the feature evaluation result and the feature interval to the comprehensive analysis module;

[0012] The comprehensive analysis module is used to implement comprehensive judgment based on the feature evaluation results and feature intervals, and decide whether to promote to the user based on the comprehensive judgment results. If promotion is to be carried out, the promotion information will be promoted according to the sequence to be promoted, and the promotion information will be a display page.

[0013] The method for generating the sequence to be promoted is:

[0014] Obtain the user's search keywords and the types of tags clicked, input the search keywords and tag types into a database for query, generate a first promotion sequence based on the query results obtained from the keyword or tag query, and in the first promotion sequence, generate a first sort from large to small by sorting application load, and generate a second sort from high to low by sorting application scores to obtain a sequence to be promoted.

[0015] The determination method of the interaction status unit is:

[0016] Get the current user's regular click count, click jump count and search box status, and make a first judgment on the search box status. If the search box is in use, the interaction status is judged to be busy. If the search box is empty, calculate the ratio of the click jump count to the regular click count to obtain the effective jump rate YT, and perform data analysis on the effective jump rate YT. If the effective jump rate YT is less than the invalid jump threshold TY within the judgment period, the interaction status is judged to be idle. Otherwise, the interaction status is judged to be busy.

[0017] The method for obtaining the interaction information is:

[0018] Monitor whether the display page of the promotion information is closed. If the display page of the promotion information is closed, generate a continuous invalidation instruction; if the display page of the promotion information is opened, generate a control instruction;

[0019] According to the continuous invalid instruction, wait for the next promotion according to the comprehensive analysis module. According to the control instruction, mark the moment when the display page is fully opened as the first starting time point, and the moment when the display page is fully closed as the end time point. Count the duration from the first starting time point to the end time point and set it as the detection interaction duration J, where the unit of J is seconds.

[0020] The method for interactive feature processing is:

[0021] The detected interaction duration J is compared with the standard interaction duration range (Jmin, Jmax). If the detected interaction duration J belongs to the standard interaction duration range (Jmin, Jmax), the detected interaction duration J is marked as the target interaction duration Ji; i represents the number of the target interaction duration, i is a positive integer, i=1, 2, 3, ..., n1; n1 represents the total number of target interaction durations;

[0022] If the detected interaction duration J is greater than the maximum value of the standard interaction duration range (Jmin, Jmax), mark the maximum value of the standard interaction duration range (Jmin, Jmax) as the end time on the timestamp as the second starting time point, and obtain the interaction activity HY within the detection period with the second starting time point as the starting point.

[0023] The method for obtaining the interactive activity HY is:

[0024] The number of jumps n2 of clicking the detail link in the display page of the promotion information is obtained during the detection period. The activity discrimination model is established based on the number of jumps n2. The expression of the activity discrimination model is:

[0025]

[0026] If the interactive activity HY is greater than 1, the target activity HY and the corresponding detection interaction duration J are marked as target activity HYj and target interaction duration Ji respectively; otherwise, the interactive activity HY and the corresponding detection interaction duration J are not marked; j represents the number of the target activity HY, j is a positive integer, j = 1, 2, ..., n3, n3 represents the total number of target activities;

[0027] An interaction feature data set is generated according to the target interaction duration Ji, the total target interaction duration n1, the target activity HYj and the total target activity n3, and the interaction feature data set is input into the standard feature model to obtain the feature evaluation value TZ.

[0028] The expression of the standard characteristic model is:

[0029]

[0030] Jc is the median of the standard interaction duration range (Jmin, Jmax); N is the actual number of promotions;

[0031] Among them, the interactive feature data set is updated after each promotion, and the feature evaluation value TZ is calculated based on the interactive feature data set obtained by the total number of recent continuous promotions;

[0032] The construction method of the personal interaction training model is:

[0033] Get the feature evaluation value TZ and establish the feature range interval The feature range interval is divided into the first feature interval (1, A], the second feature interval (A, B], and the third feature interval A and B are the one-third and two-thirds points of the characteristic range respectively.

[0034] The comprehensive judgment method is:

[0035] If the user's interaction status is idle and the feature evaluation value TZ is within the first feature interval, the user will not be promoted;

[0036] If the user's interaction status is idle and the feature evaluation value TZ is in the second feature interval or the third feature interval, the user is promoted;

[0037] If the user's interaction status is busy and the feature evaluation value TZ is within the third feature interval, the user is promoted; otherwise, the user is not promoted.

[0038] Compared with the prior art, the advantages of the present invention are:

[0039] The user's demand information is obtained through the acquisition module, and the application information in the server is matched with the demand information through the demand matching module to obtain the information matching result. A sequence to be promoted is generated according to the information matching result, and users are promoted according to the application sequence in the sequence to be promoted, making the promotion process efficient at the big data level.

[0040] The user's status is determined to be idle or busy through the interaction judgment unit, the interaction information is collected through the interaction collection unit, and the collected interaction information is analyzed through the interaction analysis module to obtain the target interaction duration, the total target interaction duration, the target activity and the total target activity, and generate an interaction feature data set. The feature evaluation value TZ is obtained according to the standard feature model, which is convenient for analyzing the user's interaction characteristics.

[0041] The personal interaction training model is constructed through the feature evaluation value TZ and the first feature interval, the second feature interval and the third feature interval are generated. The comprehensive analysis module combines the interaction status and different feature intervals for comprehensive judgment to obtain the judgment result. The promotion is carried out based on the judgment result, which optimizes the server-to-end promotion method and improves the intelligent promotion effect. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 It is a schematic diagram of the system structure of the present invention. DETAILED DESCRIPTION

[0043] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0044] In the description of the present invention, it is necessary to understand that the terms "upper", "lower", "front", "back", "left", "right", "top", "bottom", "inside", "outside", etc., indicating the orientation or position relationship are based on the orientation or position relationship shown in the drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present invention.

[0045] Reference Figure 1 , an intelligent optimization system for application promotion services based on data analysis, including a demand acquisition module, a demand matching module, an interaction collection module, an interaction analysis module and a comprehensive analysis module;

[0046] The demand acquisition module is used to obtain the user's demand information and transmit it to the application data analysis module. The demand information includes search keywords and clicked tag types. The search keywords are obtained through the search box, and the tag types are summary information of the application, such as audio and video, shopping, social networking, etc.

[0047] The demand matching module is used to obtain application information in the server, including application ratings and application loads, match the application information with the demand information, obtain information matching results, generate a sequence to be promoted based on the information matching results, and upload it to the comprehensive analysis module;

[0048] The method for generating the sequence to be promoted is:

[0049] Obtain the user's search keyword and the type of tag clicked, input the search keyword and the type of tag into the database for query, generate a first promotion sequence based on the query results obtained by the keyword or tag query, and in the first promotion sequence, generate a first sorting by sorting the application load from large to small, and generate a second sorting by sorting the application score from high to low to obtain a sequence to be promoted; the database contains application names, application tag types, application scores, and application loads; the number of applications in the sequence to be promoted can be set to 10;

[0050] An interaction collection module, including an interaction state judgment unit and an information collection unit;

[0051] The interaction status determination unit is used to determine whether the user's interaction status is idle or busy;

[0052] The information collection unit is used to obtain the user's interactive information and transmit it to the comprehensive analysis module;

[0053] The judgment method of the interactive state unit is:

[0054] Obtain the current user's regular click count, click jump count, and search box status, and make a first judgment on the search box status. If the search box is in use, the interaction status is judged to be busy; if the search box is empty, the click jump count is calculated by the ratio of the regular click count to obtain the effective jump rate YT, and perform data analysis on the effective jump rate YT. If the effective jump rate YT is less than the invalid jump threshold TY within the judgment period, the interaction status is judged to be idle; otherwise, the interaction status is judged to be busy;

[0055] The method for obtaining interactive information is:

[0056] Monitor whether the display page of the promotion information is closed. If the display page of the promotion information is closed, generate a continuous invalidation instruction; if the display page of the promotion information is opened, generate a control instruction;

[0057] According to the continuous invalid instruction, the comprehensive analysis module waits for the next promotion. According to the control instruction, the moment when the display page is fully opened is marked as the first starting time point, and the moment when the display page is fully closed is marked as the ending time point. The duration between the first starting time point and the ending time point is counted and set as the detection interaction duration J, where the unit of J is seconds;

[0058] It should be noted that the promotion information is displayed as a thumbnail box of the display page, which requires the user to click the confirmation mark to open it and click the close button to close it. Clicking close directly without clicking the confirmation button is considered a promotion failure; the thumbnail box, after being expanded, includes the application's icon information, function introduction, version information, UI display, demonstration and usage introduction video, etc., and the specific information can be jumped by clicking the link on the display page;

[0059] Among them, the opening and closing states, the start time point and the end time point of the display page are all realized by obtaining and analyzing the log, which is the existing technology and will not be described in detail;

[0060] The interaction analysis module obtains the interaction status obtained in the interaction acquisition module, and performs interaction feature processing on the interaction information to obtain an interaction feature data set, inputs the interaction feature data set into the standard feature model for feature evaluation, obtains a feature evaluation result, generates a personal interaction training model according to the feature evaluation result and the interaction status, defines a feature interval through the personal interaction training model, and transmits the feature evaluation result and the feature interval to the comprehensive analysis module;

[0061] The method of interactive feature processing is:

[0062] The detected interaction duration J is compared with the standard interaction duration range (Jmin, Jmax). If the detected interaction duration J belongs to the standard interaction duration range (Jmin, Jmax), the detected interaction duration J is marked as the target interaction duration Ji; i represents the number of the target interaction duration, i is a positive integer, i=1, 2, 3, ..., n1; n1 represents the total number of target interaction durations;

[0063] If the detected interaction duration J is greater than the maximum value of the standard interaction duration range (Jmin, Jmax), the maximum value of the standard interaction duration range (Jmin, Jmax) is marked as the end time on the timestamp as the second starting time point, and the interaction activity HY within the detection period starting from the second starting time point is obtained; the detection period can be set to 5 minutes;

[0064] The method for obtaining interactive activity HY is as follows:

[0065] The number of jumps n2 of clicking the detail link in the display page of the promotion information is obtained during the detection period. The activity discrimination model is established based on the number of jumps n2. The expression of the activity discrimination model is:

[0066]

[0067] If the interactive activity HY is greater than 1, the target activity HY and the corresponding detection interaction duration J are marked as target activity HYj and target interaction duration Ji respectively; otherwise, the interactive activity HY and the corresponding detection interaction duration J are not marked; j represents the number of the target activity HY, j is a positive integer, j = 1, 2, ..., n3, n3 represents the total number of target activities;

[0068] Generate an interactive feature data set according to the target interaction duration Ji, the total target interaction duration n1, the target activity HYj and the total target activity n3, and input the interactive feature data set into the standard feature model to obtain the feature evaluation value TZ;

[0069] The expression of the standard characteristic model is:

[0070]

[0071] Jc is the median of the standard interaction duration range (Jmin, Jmax); N is the actual number of promotions;

[0072] Among them, the interactive feature data set is updated after each promotion, and the feature evaluation value TZ is calculated based on the interactive feature data set obtained by the total number of recent continuous promotions;

[0073] The construction method of the personal interaction training model is:

[0074] Get the feature evaluation value TZ and establish the feature range interval The feature range interval is divided into the first feature interval (1, A], the second feature interval (A, B], and the third feature interval

[0075] Among them, the upper and lower limits of the feature range are both ideal conditions, that is, when the total target interaction time n1 is 0, the feature evaluation value TZ is 1; when the total target interaction time n1 is equal to the actual number of promotions and the interaction activity HYj is 2.5, the feature evaluation value TZ is Right now A and B are the one-third point and two-thirds point of the feature range interval respectively; it should be noted that ] means that the boundary value of the interval can be taken to the right, and [ means that the boundary value of the interval can be taken to the left;

[0076] The comprehensive analysis module is used to obtain the sequence to be promoted, make a comprehensive judgment based on the feature evaluation results and feature intervals, and determine whether to promote to the user based on the comprehensive judgment results. If promotion is to be carried out, the promotion information will be promoted according to the sequence to be promoted, and the promotion information is a display page;

[0077] The comprehensive judgment method is:

[0078] If the user's interaction status is idle and the feature evaluation value TZ is within the first feature interval, no promotion will be made to the user; the feature evaluation value TZ can be understood as the user's concern value for the promotion information. If the user's feature evaluation value TZ is within the first feature interval, it indicates that the user's concern value is low and the possibility of the promotion information being viewed is low, so the user will not be promoted even in the idle state;

[0079] If the user's interaction status is idle and the feature evaluation value TZ is in the second feature interval or the third feature interval, the user is promoted; if the user's feature evaluation value TZ is in the second feature interval or the third feature interval, it indicates that the user's care value is not low, and the promotion information is likely to be viewed, so the user is promoted;

[0080] If the user's interaction status is busy and the feature evaluation value TZ is within the third feature interval, the user is promoted; otherwise, the user is not promoted; if the user's feature evaluation value TZ is within the third feature interval, it indicates that the user's care value is high and the promotion information is likely to be viewed, so the user is still promoted even in the busy state; otherwise, the user's care value is not high and the user is unlikely to view the promotion information in the busy state;

[0081] By cross-analyzing and judging the feature evaluation value TZ and the interaction status, the possibility of users viewing the promotion information is increased, and the effect of the intelligent recommendation of the application is improved.

[0082] It is further explained that the above-mentioned fixed connection should be understood in a broad sense unless otherwise clearly specified and limited. For example, it can be welding, gluing, or one-piece molding, etc., which are conventional means well known to those skilled in the art.

[0083] The above description is only a preferred specific implementation manner of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can make equivalent replacements or changes according to the technical scheme and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An intelligent optimization system for application promotion services based on data analysis, characterized in that: It includes demand acquisition module, demand matching module, interaction collection module, interaction analysis module and comprehensive analysis module; The demand acquisition module is used to obtain the user's demand information and transmit it to the application data analysis module. The demand information includes search keywords and clicked tag types; the search keywords are obtained through the search box; The demand matching module is used to obtain application information in the server, including application ratings and application loads, match the application information with the demand information, obtain information matching results, generate a sequence to be promoted based on the information matching results, and upload it to the comprehensive analysis module; An interaction collection module, including an interaction state judgment unit and an information collection unit; The interaction status determination unit is used to determine whether the user's interaction status is idle or busy; The information collection unit is used to obtain the user's interactive information and transmit it to the comprehensive analysis module; The interaction analysis module obtains the interaction status obtained in the interaction acquisition module, and performs interaction feature processing on the interaction information to obtain an interaction feature data set, inputs the interaction feature data set into the standard feature model for feature evaluation, obtains a feature evaluation result, generates a personal interaction training model according to the feature evaluation result and the interaction status, defines a feature interval through the personal interaction training model, and transmits the feature evaluation result and the feature interval to the comprehensive analysis module; The comprehensive analysis module is used to implement comprehensive judgment based on the feature evaluation results and feature intervals, and decide whether to promote to the user based on the comprehensive judgment results. If promotion is to be carried out, the promotion information will be promoted according to the sequence to be promoted, and the promotion information will be a display page.

2. The intelligent optimization system for application promotion services based on data analysis according to claim 1, characterized in that: The method for generating the sequence to be promoted is: Obtain the user's search keywords and the types of tags clicked, input the search keywords and tag types into a database for query, generate a first promotion sequence based on the query results obtained from the keyword or tag query, and in the first promotion sequence, generate a first sort from large to small by sorting application load, and generate a second sort from high to low by sorting application scores to obtain a sequence to be promoted.

3. The intelligent optimization system for application promotion services based on data analysis according to claim 1 is characterized in that: The determination method of the interaction status unit is: Obtain the current user's regular click count, click jump count, and search box status, and make a first judgment on the search box status. If the search box is in use, determine that the interaction status is busy; If the search box is empty, the ratio of the number of click jumps to the number of regular clicks is calculated to obtain the effective jump rate YT, and data analysis is performed on the effective jump rate YT. If the effective jump rate YT is less than the invalid jump threshold TY within the judgment period, the interaction state is judged to be idle; otherwise, the interaction state is judged to be busy.

4. The intelligent optimization system for application promotion services based on data analysis according to claim 1, characterized in that: The method for obtaining the interaction information is: Monitor whether the display page of the promotion information is closed. If the display page of the promotion information is closed, generate a continuous invalidation instruction; if the display page of the promotion information is opened, generate a control instruction; According to the continuous invalid instruction, wait for the next promotion according to the comprehensive analysis module. According to the control instruction, mark the moment when the display page is fully opened as the first starting time point, and the moment when the display page is fully closed as the end time point. Count the duration from the first starting time point to the end time point and set it as the detection interaction duration J, where the unit of J is seconds.

5. The intelligent optimization system for application promotion services based on data analysis according to claim 4 is characterized in that: The method for processing the interactive features is: The detected interaction duration J is compared with the standard interaction duration range (Jmin, Jmax). If the detected interaction duration J belongs to the standard interaction duration range (Jmin, Jmax), the detected interaction duration J is marked as the target interaction duration Ji; i represents the number of the target interaction duration, i is a positive integer, i=1, 2, 3, ..., n1; n1 represents the total number of target interaction durations; If the detected interaction duration J is greater than the maximum value of the standard interaction duration range (Jmin, Jmax), mark the maximum value of the standard interaction duration range (Jmin, Jmax) as the end time on the timestamp as the second starting time point, and obtain the interaction activity HY within the detection period with the second starting time point as the starting point.

6. The application promotion service intelligent optimization system based on data analysis according to claim 5 is characterized by: The method for obtaining the interactive activity HY is: The number of jumps n2 of clicking the detail link in the display page of the promotion information is obtained during the detection period. The activity discrimination model is established based on the number of jumps n2. The expression of the activity discrimination model is: If the interaction activity HY is greater than 1, the target activity HY and the corresponding detection interaction duration J are marked as target activity HYj and target interaction duration Ji respectively; otherwise, the interaction activity HY and the corresponding detection interaction duration J are not marked; j represents the number of the target activity HY, j is a positive integer, j = 1, 2, ..., n3, n3 represents the total number of target activities; An interaction feature data set is generated according to the target interaction duration Ji, the total target interaction duration n1, the target activity HYj and the total target activity n3, and the interaction feature data set is input into the standard feature model to obtain the feature evaluation value TZ.

7. The application promotion service intelligent optimization system based on data analysis according to claim 6 is characterized by: The expression of the standard characteristic model is: Jc is the median of the standard interaction duration range (Jmin, Jmax); N is the actual number of promotions; Among them, the interactive feature data set is updated after each promotion, and the feature evaluation value TZ is calculated based on the interactive feature data set obtained by the total number of recent continuous promotions; The construction method of the personal interaction training model is: Get the feature evaluation value TZ and establish the feature range interval The feature range interval is divided into the first feature interval (1, A], the second feature interval (A, B], and the third feature interval A and B are the one-third and two-thirds points of the characteristic range respectively.

8. The application promotion service intelligent optimization system based on data analysis according to claim 7 is characterized by: The comprehensive judgment method is: If the user's interaction status is idle and the feature evaluation value TZ is within the first feature interval, the user will not be promoted; If the user's interaction status is idle and the feature evaluation value TZ is in the second feature interval or the third feature interval, the user is promoted; If the user's interaction status is busy and the feature evaluation value TZ is within the third feature interval, the user is promoted; otherwise, the user is not promoted.

Images