Intelligent notification pushing method and system based on applet ecology

Abstract

The invention discloses an intelligent notification pushing method and system based on applet ecology, relates to the technical field of information pushing, and solves the technical problem that an existing applet pushing system is difficult to perform information pushing according to the real-time state of a user, so that the satisfaction degree of the user to applet pushing is relatively low. The data acquisition module is used for acquiring user data in real time through a user terminal; the data processing module is used for acquiring the state of the applet when the applet is in a running state; extracting input information and browsing information in the user data, generating a target keyword group according to the input information and the browsing information, and generating push content according to the target keyword group; obtaining a user state, and generating a push instruction according to the user state and the applet state; the message pushing module is used for packaging the pushing content into pushing information; pushing the push information according to a push instruction; the content pushed by the applet is more suitable for the corresponding user; and the satisfaction of the user is enhanced.

Description

Technical Field

[0001] The present application belongs to the field of information push technology, and specifically relates to an intelligent push notification method and system based on the mini-program ecosystem. Background Art

[0002] Mini Programs are applications that can be used without downloading and installing. As a new way to connect users and services, mini programs can be easily obtained and spread within the host application and have an excellent user experience. Mini Programs include WeChat Mini Programs, Alipay Mini Programs, and Baidu Mini Programs. Mini Program developers can provide convenient and rich services within the mini programs, such as reservations, product purchases, games, and information inquiries.

[0003] Current mini program message push notifications are often randomly pushed, or some fixed messages are pushed regularly, and most of these messages are useless to users. Users frequently receive messages that do not match their current focus or actual needs, causing information overload and directly affecting user experience. At the same time, these irrelevant messages are not only difficult to stimulate users' reading interest, but may also arouse users' disgust and even cause users to choose to turn off message notifications, thereby weakening the effective interaction and stickiness between the mini program and users, resulting in low user satisfaction with the mini program push notifications. Therefore, a smart push notification method and system based on the mini program ecosystem is needed. Summary of the invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art; to this end, the present application proposes an intelligent push notification method and system based on the mini-program ecosystem, which is used to solve the technical problem that the existing mini-program push system is difficult to push information according to the user's real-time status, resulting in low user satisfaction with the mini-program push.

[0005] To achieve the above-mentioned purpose, the first aspect of the present application provides an intelligent push notification system based on the mini-program ecosystem, including: a data collection module, a data processing module, a message push module, a push storage module and a storage module;

[0006] The data acquisition module is used to acquire user data in real time through the user terminal;

[0007] The data processing module: obtains the state of the mini program when the mini program is in the running state; extracts the input information and browsing information in the user data, generates the target keyword group according to the input information and browsing information, and generates the push content according to the target keyword group; and obtains the user state, and generates the push instruction according to the user state and the mini program state;

[0008] The message push module: obtains push content, packages the push content into push information; obtains push instructions, and pushes the push information according to the push instructions.

[0009] This application obtains the running status of the mini program to determine whether the running status of the mini program is running; when the mini program is running, obtains the user's browsing information, generates the target keyword group according to the input information and browsing information in the user's browsing information; generates push content according to the target keyword group; generates push information according to the push content; obtains the mini program status and user status, generates a push instruction according to the user status and the mini program status; pushes the push information according to the push instruction. Select the content to be pushed according to the user's current user data, push it to the user in a targeted manner, and select the appropriate time to push it according to the user status, so that the content pushed by the mini program and the timing of the push are more suitable for the corresponding user; enhances the effective interaction between the mini program and the user, thereby enhancing the user's satisfaction with the mini program push.

[0010] Preferably, generating the target keyword group according to the input information and the browsed information includes:

[0011] Extracting a number of keywords and their corresponding occurrence times from the input information, and marking the occurrence times as SC;

[0012] Extracting several browsing contents and their corresponding browsing durations from the browsing information; obtaining browsing contents whose browsing durations are greater than the standard browsing durations of their corresponding browsing contents as reference contents, extracting several keywords and their corresponding occurrence times from each reference content, and marking the occurrence times as LC;

[0013] The corresponding reference target coefficient MX of each keyword is calculated by the formula MX=α1×SC+ɑ2×LC; wherein α1 and α2 are the set weight coefficients, and α1>α2;

[0014] Integrate each keyword and its corresponding reference target coefficient into a target keyword group.

[0015] Preferably, the generating of push content according to the target keyword group includes:

[0016] Acquire a number of contents to be pushed from a push storage module connected thereto; extract a number of content keywords from the contents to be pushed; input the content keywords and keywords in the target keyword group into a correlation coefficient output model to obtain the correlation coefficient between the content keywords and each keyword in each target keyword group; the correlation coefficient output model is obtained by training an artificial intelligence model;

[0017] The push reference coefficient TCj corresponding to the content keyword is calculated by the formula TCj=∑(XGji×β×MXi); wherein β is the adjustment coefficient, and 0<β<1; j is the number of several content keywords corresponding to the content to be pushed; i is the number of several keywords in the target keyword group, and i=1, 2, ..., N; N is the total number of keywords in the target keyword group; XGji is the correlation coefficient between the content keyword numbered j and the keyword numbered i;

[0018] The push coefficient TX of the corresponding content to be pushed is calculated by the formula TX=∑(TCj), where j=1, 2, ..., M; M is the total number of content keywords in the corresponding content to be pushed;

[0019] Obtain push coefficients of several contents to be pushed; and sort the contents to be pushed from large to small according to the push coefficients; determine whether the first-ranked content to be pushed is the browsing content of the browsing information; if yes, determine whether the next content to be pushed is the browsing content of the browsing information; if no, when the push coefficient corresponding to the content to be pushed is greater than the set push coefficient threshold; mark the content to be pushed as push content; otherwise, mark the push content as empty.

[0020] Preferably, the correlation coefficient output model is obtained by training an artificial intelligence model, including:

[0021] Acquire a plurality of keywords and a correlation coefficient between two keywords through a storage module; integrate the plurality of keywords and the correlation coefficient between two keywords into a plurality of training data and test data; the training data and the test data have the same data type, both including a target keyword and a plurality of comparison keywords, and a plurality of correlation coefficients; the plurality of correlation coefficients correspond to a plurality of comparison keywords one by one; the target keyword and the comparison keyword are both keywords;

[0022] The artificial intelligence model is trained using the training data, and the trained artificial intelligence model is tested using the test data, and finally a correlation coefficient output model is obtained, which inputs are the target keyword and multiple comparison keywords, and outputs are the corresponding multiple correlation coefficients;

[0023] Among them, when used, the content keyword is used as the target keyword, and several keywords in the target keyword group are used as comparison keywords to input the correlation coefficient output model; the artificial intelligence model includes the BP neural network model and the RBF neural network model.

[0024] Preferably, the step of generating a push instruction according to the user status and the mini-program status includes:

[0025] Obtain user status, which includes query status and browsing status; obtain mini-program status, which includes running status and offline status, and running status includes open status and background status; determine whether the mini-program status is open;

[0026] If yes, then when the user status is in the query status, a mini program push instruction is generated; otherwise, when the browsing is finished, a mini program push instruction is generated;

[0027] If not, when the user status is the query status, a message push instruction is generated; otherwise, when the browsing is finished, a message push instruction is generated; the push instruction includes a mini-program push instruction and a message push instruction.

[0028] This application enables the push content to be pushed to the user in a timely manner according to the status of the mini program and the selected push method; at the same time, the timing of the push is selected according to the user status, ensuring that the push content will not affect the user experience when it is pushed again.

[0029] Preferably, packaging the push content into push information comprises the following steps:

[0030] Obtain push content and push template, re-edit the push content according to the push template, and generate push information.

[0031] Another aspect of the present application provides a smart push notification method based on the mini-program ecosystem, comprising the following steps:

[0032] Step 1: Get the running status of the mini program and determine whether the running status of the mini program is running; if yes, proceed to step 2; if not, proceed to step 1;

[0033] Step 2: Obtain user browsing information, and generate target keyword groups based on input information and browsing information in the user browsing information;

[0034] Step 3: Generate push content based on target keyword groups; Generate push information based on push content;

[0035] Step 4: Obtain the mini program status and user status, and generate a push instruction based on the user status and mini program status;

[0036] Step 5: Push the push information according to the push instruction.

[0037] Compared with the prior art, the beneficial effects of this application are:

[0038] 1. This application obtains the running status of the mini program to determine whether the running status of the mini program is running; when the mini program is running, obtains the user's browsing information, generates the target keyword group according to the input information and browsing information in the user's browsing information; generates push content according to the target keyword group; generates push information according to the push content; obtains the mini program status and user status, generates a push instruction according to the user status and the mini program status; pushes the push information according to the push instruction. Select the content to be pushed according to the user's current user data, push it to the user in a targeted manner, and select the appropriate time to push it according to the user status, so that the content pushed by the mini program and the timing of the push are more suitable for the corresponding user; enhances the effective interaction between the mini program and the user, thereby enhancing the user's satisfaction with the mini program push.

[0039] 2. This application selects the push method based on the status of the mini program so that the push content can be pushed to the user in a timely manner; at the same time, the timing of pushing is selected according to the user status, ensuring that the push content will not affect the user experience when it is pushed again. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is a schematic diagram of the principle of the intelligent push notification system in this application;

[0042] Figure 2 This is a schematic diagram of the steps of the smart push notification method in this application. DETAILED DESCRIPTION

[0043] The technical solution of the present application will be described clearly and completely below in conjunction with the embodiments. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0044] See also Figure 1 , the first aspect of the present application provides an intelligent push notification system based on the mini-program ecosystem, including: a data collection module, a data processing module, a message push module, a push storage module and a storage module;

[0045] Data acquisition module: acquires user data in real time through the user terminal; user data includes input information and browsing information of the user terminal, and the input information is the information input by the user through the user terminal, including information input by chat, information input by query, etc.; user data is user data within a set time before the current time, and can be user data within 10 minutes and 30 minutes before the current time; it can be understood that when the user opens the applet corresponding to the present invention, there will be a pop-up window to remind the user whether to authorize the applet. If authorized, subsequent operations will be performed, so all data acquired in this embodiment are authorized by the user;

[0046] Data processing module: obtain the mini program status, the mini program status is the current status of the mini program, including the running status and offline status. The running status means that the mini program is running, including the open status and background status. The open status means that the user can check the messages pushed through the mini program in time; the background status means that the user cannot check the content recommended by the mini program in time; the offline status means that the mini program cannot be pushed at present; when the mini program is in the running state; extract the input information and browsing information in the user data, generate the target keyword group according to the input information and browsing information, the target keyword group is the keyword combination of the content that the user wants to query according to the user data; generate the push content according to the target keyword group; the push content is the content that needs to be pushed this time; and obtain the user status, the user status includes the query status and browsing status; the query status is the status when the user is inputting information, and the browsing status is the status when the user is browsing content; generate the push instruction according to the user status and the mini program status; the push instruction is the instruction to control the push of push information;

[0047] In this embodiment, the user status and applet status are stored in the Redis database in the storage module, while the user data storage and feedback data are stored in the MySQL database in the storage module;

[0048] Message push module: obtain push content and package the push content into push information; the push information is information obtained by reintegrating the push content according to a fixed push format; obtain push instructions and push the push information according to the push instructions.

[0049] This embodiment obtains the running status of the mini program to determine whether the running status of the mini program is in the running status; when the mini program is in the running status, obtains the user's browsing information, generates the target keyword group according to the input information and browsing information in the user's browsing information; generates the push content according to the target keyword group; generates the push information according to the push content; obtains the mini program status and the user status, generates the push instruction according to the user status and the mini program status; pushes the push information according to the push instruction. The content to be pushed is selected according to the user's current user data, and the user is pushed in a targeted manner. At the same time, according to the user status, the appropriate time is selected for pushing, so that the content pushed by the mini program and the time of pushing are more suitable for the corresponding user; the effective interaction between the mini program and the user is enhanced, thereby enhancing the user's satisfaction with the mini program push.

[0050] Generating a target keyword group according to the input information and the browsing information includes: extracting a number of keywords and their corresponding occurrence times in the input information, and marking the occurrence times as SC;

[0051] Extract several browsing contents and their corresponding browsing durations in the browsing information; the browsing contents are the contents corresponding to the web pages, documents and messages browsed by the user, and the browsing duration is the duration that the user stays on the corresponding browsing contents (which can be further extended); obtain browsing contents whose browsing duration is longer than the standard browsing duration of the corresponding browsing contents as reference contents, and the standard browsing duration is obtained in the following manner; obtain several historical browsing contents and their corresponding historical browsing durations; group the historical browsing contents according to the number of words in the contents, such as grouping them at intervals of 100 words, with 0-100 words being grouped into one group and 100-200 words being grouped into another group, etc.; obtain the historical browsing durations corresponding to the historical browsing contents in each group, and use the average value of each browsing duration as the standard browsing duration of the corresponding group; query the standard browsing duration of the corresponding group according to the number of words corresponding to the current browsing contents.

[0052] Extracting a number of keywords and their corresponding occurrence times from each reference content, and marking the occurrence times as LC;

[0053] The corresponding reference target coefficient MX of each keyword is calculated by the formula MX=α1×SC+ɑ2×LC; wherein α1 and α2 are the set weight coefficients, and α1>α2; the keywords in the input information are more reference-oriented than the keywords in the browsed content; therefore, the number of keywords in the input information has a larger weight coefficient than the number of keywords in the browsed content, that is, α1>α2;

[0054] Integrate each keyword and its corresponding reference target coefficient into a target keyword group.

[0055] This embodiment calculates the reference target coefficient of each keyword through the above formula, and generates its corresponding reference target coefficient by referring to the current user data and the number of times the keyword appears in the input information and browsed content in the user data. The more times the keyword appears, the greater its importance in the current user query process, and the larger its corresponding reference target coefficient.

[0056] Generate push content based on target keyword groups, including:

[0057] Acquire a number of contents to be pushed from a push storage module connected thereto; extract a number of content keywords from the contents to be pushed; input the content keywords and keywords in the target keyword group into a correlation coefficient output model to obtain the correlation coefficient between the content keywords and each keyword in each target keyword group; the correlation coefficient output model is obtained by training an artificial intelligence model;

[0058] The push reference coefficient TCj corresponding to the content keyword is calculated by the formula TCj=∑(XGji×β×MXi); wherein β is the adjustment coefficient, and 0<β<1; j is the number of several content keywords corresponding to the content to be pushed; i is the number of several keywords in the target keyword group, and i=1, 2, ..., N; N is the total number of keywords in the target keyword group; XGji is the correlation coefficient between the content keyword numbered j and the keyword numbered i;

[0059] This embodiment calculates the push reference coefficient of each content keyword through the above formula; it is specifically generated based on the correlation coefficient between the content keyword and each keyword in the target keyword group, and the reference target coefficient of each keyword; when the correlation between the content keyword and each keyword in the target keyword group is greater, and the reference target coefficient of each keyword is greater, the matching degree between its corresponding keyword and the target keyword group is higher, and the corresponding push reference coefficient is greater.

[0060] The push coefficient TX of the corresponding content to be pushed is calculated by the formula TX=∑(TCj), where j=1, 2, ..., M; M is the total number of content keywords in the corresponding content to be pushed;

[0061] This embodiment calculates the recommendation coefficient of each content to be pushed through the above formula. When the push reference coefficient corresponding to each content keyword in the content to be pushed is larger, the corresponding content to be pushed is more in line with the user's current interests, and therefore the corresponding recommendation coefficient value is larger.

[0062] Obtain push coefficients of several contents to be pushed; and sort the contents to be pushed from large to small according to the push coefficients; determine whether the first-ranked content to be pushed is the browsing content of the browsing information; if yes, determine whether the next content to be pushed is the browsing content of the browsing information; if not, when the push coefficient corresponding to the content to be pushed is greater than the set push coefficient threshold; mark the content to be pushed as push content; otherwise, mark the push content as empty; this embodiment can recommend the content to be pushed by excluding the content that the user has browsed, or directly generate the push content from the first-ranked content to be pushed.

[0063] The correlation coefficient output model is obtained through artificial intelligence model training, including:

[0064] A plurality of keywords and a correlation coefficient between two keywords are obtained through a storage module; the correlation coefficient can be obtained through its corresponding cosine similarity, or can be set by an expert; a plurality of keywords and a correlation coefficient between two keywords are integrated into a plurality of training data and test data; the data types of the training data and the test data are the same, both including a target keyword and a plurality of comparison keywords, and a plurality of correlation coefficients; the plurality of correlation coefficients correspond to a plurality of comparison keywords one by one; the target keyword and the comparison keyword are both keywords; when the target keyword and the comparison keyword are the same, the corresponding correlation coefficient is "1"; specifically, the correlation coefficient among the plurality of correlation coefficients is the correlation coefficient between the target keyword and the comparison keyword among the plurality of comparison keywords;

[0065] The artificial intelligence model is trained using the training data, and the trained artificial intelligence model is tested using the test data. Specifically, the target keyword and multiple comparison keywords in the test data are input into the trained artificial intelligence model to obtain multiple corresponding correlation coefficients. When the difference between the output correlation coefficient and the corresponding correlation coefficient recorded in the test data is within an acceptable range, the next set of test data is tested; otherwise, the relevant parameters are adjusted until the output result of this set of test data is within an acceptable range, and until the proportion of the test data within the acceptable range is greater than the set ratio, a correlation coefficient output model is obtained, which has the target keyword and multiple comparison keywords as input and the corresponding multiple correlation coefficients as output;

[0066] Among them, when used, the content keyword is used as the target keyword, and several keywords in the target keyword group are used as comparison keywords to input the correlation coefficient output model; the artificial intelligence model includes the BP neural network model and the RBF neural network model.

[0067] It is worth noting that in this embodiment, the push is also evaluated through the feedback data, specifically including: extracting the user browsing time corresponding to the push content in the feedback data and the corresponding standard browsing time, when the user browsing time is less than the standard browsing time, the evaluation of the push is set to failure; when the user browsing time is longer than the standard browsing time, the evaluation of the push is set to success; and adjusting the content and timing of the push according to the evaluation;

[0068] Generate a push instruction according to the user status and the mini program status, including: obtaining the user status, which includes the query status and the browsing status; obtaining the mini program status, which includes the running status and the offline status, and the running status includes the open status and the background status; and determining whether the mini program status is the open status;

[0069] If yes, then when the user status is in the query status, a mini program push instruction is generated; otherwise, when the browsing is finished, a mini program push instruction is generated;

[0070] If not, when the user status is the query status, a message push instruction is generated; otherwise, when the browsing is finished, a message push instruction is generated; the push instruction includes a mini-program push instruction and a message push instruction.

[0071] This embodiment enables the push content to be pushed to the user in a timely manner according to the status of the mini-program and the selected push method; at the same time, the timing of the push is selected according to the user status, ensuring that the push content will not affect the user experience when it is pushed again.

[0072] Packing the push content into push information includes the following steps: acquiring the push content and a push template, re-editing the push content according to the push template, and generating push information.

[0073] See also Figure 2 Another aspect of the present application provides a smart push notification method based on the mini-program ecosystem, comprising the following steps:

[0074] Step 1: Get the running status of the mini program and determine whether the running status of the mini program is running; if yes, proceed to step 2; if not, proceed to step 1;

[0075] Step 2: Obtain user browsing information, and generate target keyword groups based on input information and browsing information in the user browsing information;

[0076] Step 3: Generate push content based on target keyword groups; Generate push information based on push content;

[0077] Step 4: Obtain the mini program status and user status, and generate a push instruction based on the user status and mini program status;

[0078] Step 5: Push the push information according to the push instruction.

[0079] Part of the data in the above formula is calculated by removing the dimension and taking its numerical value. The formula is a formula closest to the actual situation obtained by software simulation of a large amount of collected data; the preset parameters and preset thresholds in the formula are set by technical personnel in this field according to actual conditions or obtained through simulation of a large amount of data.

[0080] How this application works:

[0081] This application obtains the running status of the mini program to determine whether the running status of the mini program is running; when the mini program is running, obtains the user's browsing information, generates the target keyword group according to the input information and browsing information in the user's browsing information; generates push content according to the target keyword group; generates push information according to the push content; obtains the mini program status and user status, generates a push instruction according to the user status and the mini program status; pushes the push information according to the push instruction. Select the content to be pushed according to the user's current user data, push it to the user in a targeted manner, and select the appropriate time to push it according to the user status, so that the content pushed by the mini program and the timing of the push are more suitable for the corresponding user; enhances the effective interaction between the mini program and the user, thereby enhancing the user's satisfaction with the mini program push.

[0082] The above embodiments are only used to illustrate the technical method of the present application and are not intended to limit it. Although the present application has been described in detail with reference to the preferred embodiments, a person of ordinary skill in the art should understand that the technical method of the present application may be modified or replaced by equivalents without departing from the spirit and scope of the technical method of the present application.

Claims

1. An intelligent push notification system based on the mini program ecosystem, comprising: The data acquisition module, the data processing module, the message push module, the push storage module and the storage module are characterized in that: The data acquisition module is used to acquire user data in real time through the user terminal; The data processing module: obtains the state of the mini program when the mini program is in the running state; extracts the input information and browsing information in the user data, generates the target keyword group according to the input information and browsing information, and generates the push content according to the target keyword group; and obtains the user state, and generates the push instruction according to the user state and the mini program state; The message push module: obtains push content, packages the push content into push information; obtains push instructions, and pushes the push information according to the push instructions.

2. According to claim 1, a smart push notification system based on mini-program ecology is characterized in that: The generating of the target keyword group according to the input information and the browse information includes: Extracting a number of keywords and their corresponding occurrence times from the input information, and marking the occurrence times as SC; Extracting several browsing contents and their corresponding browsing durations from the browsing information; obtaining browsing contents whose browsing durations are greater than the standard browsing durations of their corresponding browsing contents as reference contents, extracting several keywords and their corresponding occurrence times from each reference content, and marking the occurrence times as LC; The corresponding reference target coefficient MX of each keyword is calculated by the formula MX=α1×SC+ɑ2×LC; wherein α1 and α2 are the set weight coefficients, and α1>α2; Integrate each keyword and its corresponding reference target coefficient into a target keyword group.

3. According to claim 1, the intelligent push notification system based on the mini program ecosystem is characterized in that: The generating of push content according to the target keyword group includes: Acquire a number of contents to be pushed from a push storage module connected thereto; extract a number of content keywords from the contents to be pushed; input the content keywords and keywords in the target keyword group into a correlation coefficient output model to obtain the correlation coefficient between the content keywords and each keyword in each target keyword group; the correlation coefficient output model is obtained by training an artificial intelligence model; The push reference coefficient TCj corresponding to the content keyword is calculated by the formula TCj=∑(XGji×β×MXi); wherein β is the adjustment coefficient, and 0<β<1; j is the number of several content keywords corresponding to the content to be pushed; i is the number of several keywords in the target keyword group, and i=1, 2, ..., N; N is the total number of keywords in the target keyword group; XGji is the correlation coefficient between the content keyword numbered j and the keyword numbered i; The push coefficient TX of the corresponding content to be pushed is calculated by the formula TX=∑(TCj), where j=1, 2, ..., M; M is the total number of content keywords in the corresponding content to be pushed; Obtain push coefficients of several contents to be pushed; and sort the contents to be pushed from large to small according to the push coefficients; determine whether the first-ranked content to be pushed is the browsing content of the browsing information; if yes, determine whether the next content to be pushed is the browsing content of the browsing information; if no, when the push coefficient corresponding to the content to be pushed is greater than the set push coefficient threshold; mark the content to be pushed as push content; otherwise, mark the push content as empty.

4. According to claim 3, the intelligent push notification system based on the mini program ecosystem is characterized in that: The correlation coefficient output model is obtained through artificial intelligence model training, including: Acquire a plurality of keywords and a correlation coefficient between two keywords through a storage module; integrate the plurality of keywords and the correlation coefficient between two keywords into a plurality of training data and test data; the training data and the test data have the same data type, both including a target keyword and a plurality of comparison keywords, and a plurality of correlation coefficients; the plurality of correlation coefficients correspond to a plurality of comparison keywords one by one; the target keyword and the comparison keyword are both keywords; The artificial intelligence model is trained using the training data, and the trained artificial intelligence model is tested using the test data, and finally a correlation coefficient output model is obtained, which inputs are the target keyword and multiple comparison keywords, and outputs are the corresponding multiple correlation coefficients; Among them, when used, the content keyword is used as the target keyword, and several keywords in the target keyword group are used as comparison keywords to input the correlation coefficient output model; the artificial intelligence model includes the BP neural network model and the RBF neural network model.

5. According to claim 1, the intelligent push notification system based on the mini program ecosystem is characterized in that: The generating of a push instruction according to the user status and the mini-program status includes: Obtain user status, which includes query status and browsing status; obtain mini-program status, which includes running status and offline status, and running status includes open status and background status; determine whether the mini-program status is open; If yes, then when the user status is in the query status, a mini program push instruction is generated; otherwise, when the browsing is finished, a mini program push instruction is generated; If not, when the user status is the query status, a message push instruction is generated; otherwise, when the browsing is finished, a message push instruction is generated; the push instruction includes a mini-program push instruction and a message push instruction.

6. According to claim 1, a smart push notification system based on mini-program ecology is characterized in that: Packaging the push content into push information includes the following steps: Obtain push content and push template, re-edit the push content according to the push template, and generate push information.

7. A smart push notification method based on mini-program ecology, based on the operation of the smart push notification system based on mini-program ecology in any one of claims 1 to 6; characterized in that: The following steps are involved: Step 1: Get the running status of the mini program and determine whether the running status of the mini program is running; if yes, proceed to step 2; if not, proceed to step 1; Step 2: Obtain user browsing information, and generate target keyword groups based on input information and browsing information in the user browsing information; Step 3: Generate push content based on target keyword groups; Generate push information based on push content; Step 4: Obtain the mini program status and user status, and generate a push instruction based on the user status and mini program status; Step 5: Push the push information according to the push instruction.

Images