WO2022077921A1

WO2022077921A1 - Method and apparatus for pushing dynamic epidemic prevention knowledge, device, and storage medium

Info

Publication number: WO2022077921A1
Application number: PCT/CN2021/097074
Authority: WO
Inventors: 徐卓扬; 孙行智; 胡岗; 赵惟; 左磊; 赵婷婷
Original assignee: 平安科技（深圳）有限公司
Priority date: 2020-10-12
Filing date: 2021-05-30
Publication date: 2022-04-21
Also published as: CN112148937A; CN112148937B

Abstract

A method for pushing dynamic epidemic prevention knowledge, comprising: extracting, from self-examination report data, feature information of each user in a target population; acquiring an epidemic prevention guide of the latest version and an epidemic prevention guide of a previous version, screening updated epidemic prevention knowledge points which are not pushed to the target population, and taking the screened epidemic prevention knowledge points as knowledge points to be pushed; adjusting the format of feature information contained in the self-examination report data to be a format consistent with the knowledge points to be pushed, and generating an adjusted feature information set (S3); and constructing and training a push model on the basis of the self-examination report data and the knowledge points to be pushed, inputting the adjusted feature information set to the trained push model, and outputting an epidemic prevention knowledge point having the largest intervention effect on the feature information set and pushing same to a preset user end (S4). The present method can monitor, in real time, the change of self-examination report data, thereby automatically matching a suitable epidemic prevention knowledge point according to the real-time self-examination report data, and pushing same to a target user group.

Description

Push method, device, equipment and storage medium for dynamic epidemic prevention knowledge

This application claims the priority of the Chinese patent application with the application number CN202011084986.7 and the invention titled "Method and System for Pushing Dynamic Epidemic Prevention Knowledge" submitted to the Chinese Patent Office on October 12, 2020, the entire contents of which are incorporated by reference in in this application.

technical field

The present application relates to the field of big data, and in particular, to a method, device, device and storage medium for pushing dynamic epidemic prevention knowledge.

Background technique

To establish a push system for patient education, it is generally necessary to collect a large number of samples for learning. To verify a push system, prospective follow-up is required for verification. Collecting samples and conducting follow-up will require a lot of time and money, and a lot of money. Time and money to collect samples and follow-up, just for patient education push, obviously has no market value;

During special periods (such as during the epidemic), government departments, neighborhood committees and other institutions may require residents to conduct self-examination and report daily, and update daily health information on their own mobile terminals. This information has the characteristics of low consumption and easy access. Desensitization, these self-reported data can also be used for proper analysis;

However, the inventor realized that the problem of the existing patient education push system is that it lacks the means to reasonably apply the self-examination and reported data in a special period, so it takes a lot of time and money to collect data and conduct follow-up; in addition, the existing The patient education system also lacks a comprehensive dynamic update mechanism, which cannot be applied to the scenarios where the population self-examination and reporting data and the epidemic prevention knowledge points in the epidemic prevention guide change dynamically during the epidemic prevention period. Push knowledge points to target user groups. For example, during the epidemic prevention period, the epidemic prevention knowledge points and the health data of the population are "dynamic", and the epidemic prevention guide version is updated and iterated rapidly, and a certain version of the epidemic prevention knowledge pushed may be changed, updated or cancelled by future versions.

SUMMARY OF THE INVENTION

A method for pushing dynamic epidemic prevention knowledge, applied to an electronic device, the method includes:

S1, acquiring self-inspection and reporting data of a target group on a preset date from a data source, and extracting characteristic information of each user in the target group from the self-inspection and reporting data;

S2, obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, screen out the updated and unpublished epidemic prevention knowledge points to the target population, and Take the selected knowledge points for epidemic prevention as the knowledge points to be pushed;

S3, adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generating an adjusted feature information set;

S4, build and train a push model based on the self-checked and reported data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the epidemic prevention knowledge with the greatest intervention effect on the feature information set Click to push to the default client.

A push device for dynamic epidemic prevention knowledge, the device includes:

an extraction module, configured to obtain self-inspection and reporting data of the target group on a preset date from a data source, and extract characteristic information of each user in the target group from the self-inspection and reporting data;

The analysis module is used to obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, and screen out the updated and unpublished epidemic prevention knowledge to the target population point, and use the selected epidemic prevention knowledge points as the knowledge points to be pushed;

Processing module: used to adjust the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generate an adjusted feature information set; and

Pushing module: used to construct and train a push model according to the self-checking and reporting data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the maximum intervention effect on the feature information set The epidemic prevention knowledge points are pushed to the preset user terminal.

An electronic device, the electronic device includes: a memory and a processor, the memory stores a push program of dynamic epidemic prevention knowledge that can be run on the processor, and the push program of the dynamic epidemic prevention knowledge is stored by the processor The following steps are implemented when executing:

Obtain self-inspection and reporting data of the target group on a preset date from a data source, and extract characteristic information of each user in the target group from the self-inspection and reporting data;

Obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, screen out the updated and unpublished epidemic prevention knowledge points to the target population, and screen The published epidemic prevention knowledge points are used as knowledge points to be pushed;

Adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generating an adjusted feature information set;

Build and train a push model based on the self-checked and reported data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the epidemic prevention knowledge point push with the greatest intervention effect on the feature information set to the default client.

A computer-readable storage medium, the computer-readable storage medium includes a push program of dynamic epidemic prevention knowledge, and the push program of the dynamic epidemic prevention knowledge is executed by a processor to achieve the following steps:

This application can automatically match the appropriate epidemic prevention knowledge points with the real-time self-inspection and report data and push it to the target user group.

Description of drawings

1 is a schematic flowchart of a method for pushing dynamic epidemic prevention knowledge of the application;

Fig. 2 is the schematic diagram of the electronic device of the application;

Fig. 3 is the module schematic diagram of the push device of the dynamic epidemic prevention knowledge of the application;

The realization, functional characteristics and advantages of the purpose of the present application will be further described with reference to the accompanying drawings in conjunction with the embodiments.

Detailed ways

In order to make the purpose, technical solutions and advantages of the present application more clearly understood, the present application will be described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, but not to limit the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

The present application provides a method for pushing dynamic epidemic prevention knowledge. Referring to FIG. 1 , it is a schematic flowchart of a method for pushing dynamic epidemic prevention knowledge of the present application. The method may be performed by an electronic device, which may be implemented by software and/or hardware.

In this embodiment, a method for pushing dynamic epidemic prevention knowledge includes the following steps:

Specifically, the data source comes from the daily self-inspection and reporting system of the local neighborhood committee/street (during the epidemic prevention period, government departments, neighborhood committees and other institutions require residents to conduct daily self-inspection and report, and residents log in to their mobile terminals for daily self-inspection. It should be noted that, in this embodiment, the “self-inspection and reporting data of the preset date” refers to the self-inspection and reporting data of the target group as of the previous day. Check the reported data.

Further, the characteristic information includes the user's basic information, going out records, health status and the answer to the symptom question; the user's basic information includes gender, age, height and weight, etc.; the health status is "whether the disease is on"; The answers to the symptom questions are answers to questions such as "whether dizziness occurs" and "whether fatigue occurs".

More preferably, the S1 further includes: performing fuzzy processing on the sensitive information in the self-inspection and reporting data through a fuzzy algorithm to obtain fuzzy data, and replacing the self-inspection and reporting data with the fuzzy data to realize the self-inspection and reporting data. Check and report data desensitization. The sensitive information: mainly involves the user's private information, such as: residential address, ID number, work unit and other information.

In this embodiment, in order to ensure the authenticity, reliability and timeliness of epidemic prevention knowledge points, the epidemic prevention guide comes from official channels such as official websites of government departments, authoritative forums, major media news channels, and medical websites.

Specifically, a comparative analysis of the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide includes the following steps:

Recognize the text in the latest version of the epidemic prevention guide through the OCR engine and generate the first recognition text, and use the OCR engine to recognize the text in the previous version of the epidemic prevention guide and generate the second recognition text, and the first recognition text contains the recognition text. Publish the text in the latest version of the epidemic prevention guide and the coordinate information of each text in the corresponding OCR identification image, and the second identification text includes the text in the previous version of the epidemic prevention guide and the coordinate information of each character in the corresponding OCR identification image;

Use a text comparison algorithm to compare the difference text between the first recognition text and the second recognition text, and obtain the coordinate information of the difference text;

The difference text is located, and the difference text is marked in the latest version of the epidemic prevention guide.

It should be noted that, in this embodiment, the text comparison algorithm adopts a queue comparison method, and establishes a queue respectively for the first recognition text and the second recognition text to be compared, and then compares them word by word. Find the same text and different text according to the two queues, then form an identical text queue and a difference text queue. The characters include Chinese characters, English letters, numbers, symbols, etc., but are not limited thereto.

Further, before the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide are submitted to the OCR engine, preprocessing is required, and the preprocessing includes removing interference stamps, ink dots, underlines, marking headers and footers, and/or markings. Table position; also includes image enhancement and sharpening, grayscale, binarization, noise reduction, tilt correction, etc., so as to detect the seal, ink dot, and underline of interfering characters, and remove these interferences before character recognition, and then Then mark the header and footer or mark the table position. Through preprocessing, the accuracy of OCR engine recognition of text can be improved.

S3, adjusting the format of the feature information contained in the self-inspection and reporting data to be consistent with the format of the knowledge point to be pushed, and generating an adjusted feature information set;

Before performing step S3, the feature information needs to be discretized. In detail, the discretization process includes: using a dynamic programming method to find all word segmentation combinations of the feature information, and calculating each word segmentation combination The word weights below are traversed to obtain the word segmentation and combination corresponding to the word weight with the largest value, and the discrete text set of the feature information under the word segmentation and combination is obtained.

Further, the dynamic programming rule to find word segmentation combinations includes: constructing a gradient descent algorithm and an iterative function based on the feature information, solving the segmentation parameters of the iterative function, and solving the gradient descent according to different segmentation parameters. The result value of the algorithm, according to the result value, different word segmentation combinations are obtained.

Preferably, the gradient descent algorithm is:

Among them, θ is the segmentation parameter, J(θ) is the different segmentation combinations based on the θ, ^xi represents the vector representation of the i-th word of the feature information,

Represents the vector representation of the jth word of the feature information, t represents the number of words represented by the vector before the jth word, and T represents the transpose of the matrix.

The iteration function is:

where δ is the noise parameter of the iterative function.

Further, the calculation method of the word weight is:

Among them, WS(xi) represents the weight of the ith word, d is the damping coefficient, In(xi) represents the situation when no word segmentation is performed, and the feature information is the feature information when no word segmentation is performed in this application. , out(xi) represents the situation where the word segmentation and combination have been completed, TextRank(xi) represents the word criticality score calculated according to the TextRank algorithm, wi represents the i-th word in the out(xi) word segmentation and combination The proportion of occurrences below.

According to the prompt of the knowledge point to be pushed, the words in the discrete text set of the feature information are adjusted. The feature information corresponding to "age" in the discrete text set is converted into "yes/no middle-aged and elderly people"; if the knowledge point to be pushed indicates that "people with cardiovascular disease are susceptible", the discrete text set of the feature information is set. The feature information corresponding to "Is it sick" is adjusted to "Whether there is cardiovascular disease"; if the knowledge point to be pushed prompts "Reduce going out", it is adjusted according to the feature information corresponding to "Going out record" in the discrete text set of the feature information is "the number of times to go out";

Specifically, the feature information set is in the form of {"the middle-aged and the elderly are generally susceptible" → "whether the middle-aged and the elderly are", "people with cardiovascular disease are susceptible" → "whether there is cardiovascular and cerebrovascular disease", "reduce going out" → "Number of trips out"...}.

It should be noted that, in this embodiment, the push model is obtained by training based on the self-inspection and reported data and the epidemic prevention knowledge points to be pushed, and the training method includes the following steps:

Taking the self-examination and reporting data and the knowledge points to be pushed as training samples;

dividing the training sample into a training set and a verification set according to a preset ratio;

Inputting the training set into a preset training model for regression training, and using the verification set every preset period to verify the accuracy of the model; and

When the accuracy rate is greater than the preset threshold, the training is ended, and the push model is obtained.

Further, the push model formula is:

h(t,X(t))=h ₀ (t)exp(β ₁ X ₁ (t)+β ₂ X ₂ (t)+...β _m X _m (t))

Among them, h ₀ (t) is the reference rate, exp is an exponential function with the base of natural constant e, and the value of constant e is 2.718282;

X(t) represents the feature information set of the t day, X _m (t) represents the mth feature information of the t day; β _m represents the coefficient corresponding to the mth feature information of the t day;

h(t,X(t)) represents the knowledge point of epidemic prevention with the greatest intervention effect on the characteristic information set input on the t day.

In this embodiment, the working principle of the push model is as follows: after the self-inspection and reporting data is updated and/or the epidemic prevention guide is updated, all the adjusted feature information in the feature information set is input to the preset In the trained push model, the push model outputs the epidemic prevention knowledge point with the greatest intervention effect on the feature information set and pushes it to the preset user terminal. For example, after the push model pushes a certain epidemic prevention knowledge point on the t day, the push model is retrained on the t+1 day. The difference from the model on the t day is that the training feature information has been updated, and the epidemic prevention guide The epidemic prevention knowledge point of t is updated (that is, the characteristic information set changes); the characteristic information corresponding to the epidemic prevention knowledge point pushed on the t day is deleted from the input characteristic information set.

Please refer to FIG. 2 , which is a schematic diagram of the electronic device 1 of the present application.

In this embodiment, the push program 10 of dynamic epidemic prevention knowledge is installed and run in the electronic device 1 . The electronic device 1 may be a computing device such as a desktop computer, a notebook, a palmtop computer, and a server. The electronic device 1 may include, but is not limited to, a memory 11 , a processor 12 and a display 13 . Figure 2 only shows the electronic device 1 with components 11-13, but it should be understood that implementation of all shown components is not required, and that more or fewer components may be implemented instead.

Wherein, the memory 11 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static Random Access Memory (SRAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), Programmable Read Only Memory (PROM), Magnetic Memory, Magnetic Disk, Optical Disk, etc. In some embodiments, the memory 11 may be an internal storage unit of the server 1 , such as a hard disk or a memory of the server 1 . In other embodiments, the memory 11 may also be an external storage device of the server 1, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital) equipped with the server 1 , SD) card, flash memory card (Flash Card), etc. Of course, the memory 11 may also include both the internal storage unit of the server 1 and its external storage device. In this embodiment, the memory 11 is generally used to store the operating system and various application software installed on the server 1 , such as the program code of the push program 10 for dynamic epidemic prevention knowledge, and the like. In addition, the memory 11 can also be used to temporarily store various types of data that have been output or will be output.

In some embodiments, the processor 12 may be a central processing unit (Central Processing Unit, CPU), a controller, a microcontroller, a microprocessor, or other data processing chips. The processor 12 is generally used to control the overall operation of the server 1, such as performing data interaction or communication-related control and processing. In this embodiment, the processor 12 is configured to run the program code or process data stored in the memory 11 , for example, run the program code of the push program 10 for dynamic epidemic prevention knowledge, and the like.

In some embodiments, the display 13 may be an LED display, a liquid crystal display, a touch-sensitive liquid crystal display, an OLED (Organic Light-Emitting Diode, organic light-emitting diode) touch device, and the like. The display 13 is used to display the information processed in the electronic device 1 and to display a visual user interface, such as a push interface of dynamic epidemic prevention knowledge, and the like. The components 11-13 of the electronic device 1 communicate with each other via a system bus.

In the above embodiment, when the processor 12 executes the push program of the dynamic epidemic prevention knowledge stored in the memory 11, the following steps may be implemented:

Please refer to FIG. 3 , which is a schematic block diagram of the device 100 for pushing dynamic epidemic prevention knowledge of the present application. In this embodiment, the apparatus 100 for pushing dynamic epidemic prevention knowledge can be divided into one or more modules, one or more modules are stored in the memory 11, and are processed by one or more processors (in this embodiment, the processor 12) performed to complete this application. For example, in FIG. 3 , the apparatus 100 for pushing dynamic epidemic prevention knowledge can be divided into an extraction module 101 , an analysis module 102 , a processing module 103 and a push module 104 . The module referred to in this application refers to a series of computer program instruction segments capable of accomplishing specific functions, wherein:

Extraction module 101: used to obtain self-inspection and reporting data of a target group on a preset date from a data source, and extract characteristic information of each user in the target group from the self-inspection and reporting data;

More preferably, the extraction module 101 includes a desensitization unit, which is used for: fuzzing the sensitive information in the self-inspection and reporting data through a fuzzy algorithm to obtain fuzzy data, and using the fuzzy data to report the self-inspection and reporting. Data replacement implements desensitization of the self-checking and reporting data. The sensitive information: mainly involves the user's private information, such as: residential address, ID number, work unit and other information.

Analysis module 102: used to obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, and screen out the updated and unpublished epidemic prevention guidelines for the target population Knowledge points, and use the selected epidemic prevention knowledge points as knowledge points to be pushed;

Processing module 103: used to adjust the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generate an adjusted feature information set;

Preferably, the gradient descent algorithm is:

The iteration function is:

where δ is the noise parameter of the iterative function.

Further, the calculation method of the word weight is:

Pushing module 104: for constructing and training a push model according to the self-checking and reporting data and the knowledge points to be pushed, inputting the feature information set into the trained push model, and outputting the intervention effect on the feature information set The largest epidemic prevention knowledge points are pushed to the preset user terminal.

It should be noted that, in this embodiment, the push model is obtained by training based on the self-inspection and reported data and the epidemic prevention knowledge points to be pushed. Specifically, the push module 104 includes a training unit for training the push model, and the training process includes the following steps:

Further, the push model formula is:

h(t,X(t))=h ₀ (t)exp(β ₁ X ₁ (t)+β ₂ X ₂ (t)+...β _m X _m (t))

In this embodiment, the working principle of the push model is as follows: after the self-inspection and reporting data is updated and/or the epidemic prevention guide is updated, all the adjusted feature information in the feature information set is input to the preset In the trained push model, the push model outputs the epidemic prevention knowledge point with the greatest intervention effect on the feature information set and pushes it to the preset user terminal. For example, after the push model pushes a certain epidemic prevention knowledge point on the t day, the push model is retrained on the t+1 day. The difference from the model on the t day is that the training feature information has been updated, and the epidemic prevention guide The epidemic prevention knowledge point of t is updated (that is, the characteristic information set changes); the characteristic information corresponding to the epidemic prevention knowledge point pushed on the t day is deleted from the input characteristic information.

In addition, an embodiment of the present application also proposes a computer-readable storage medium, the computer-readable storage medium may be non-volatile or volatile, and the computer-readable storage medium is stored with dynamic epidemic prevention knowledge A push program, the push program of the dynamic epidemic prevention knowledge can be executed by one or more processors to achieve the following operations:

In another embodiment, in the method for pushing dynamic epidemic prevention knowledge provided by this application, in order to further ensure the privacy and security of all the above-mentioned data, all the above-mentioned data can also be stored in a node of a blockchain. For example, user feature information and knowledge points to be pushed, etc., these data can be stored in the blockchain nodes.

It should be noted that the blockchain referred to in this application is a new application mode of computer technologies such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm. Blockchain, essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information to verify its Validity of information (anti-counterfeiting) and generation of the next block. The blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

The above are only the preferred embodiments of the present application, and are not intended to limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the contents of the description and drawings of the present application, or directly or indirectly applied in other related technical fields , are similarly included within the scope of patent protection of this application.

Claims

A method for pushing dynamic epidemic prevention knowledge, applied to electronic equipment, wherein the method includes:

S1, acquiring self-inspection and reporting data of a target group on a preset date from a data source, and extracting characteristic information of each user in the target group from the self-inspection and reporting data;

S2, obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, screen out the updated and unpublished epidemic prevention knowledge points to the target population, and Take the selected knowledge points for epidemic prevention as the knowledge points to be pushed;

S3, adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generating an adjusted feature information set;

S4, build and train a push model based on the self-checked and reported data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the epidemic prevention knowledge with the greatest intervention effect on the feature information set Click to push to the default client.
The method for pushing dynamic epidemic prevention knowledge according to claim 1, wherein the characteristic information includes the user's basic information, going out records, health status, and answers to symptom questions.
The method for pushing dynamic epidemic prevention knowledge according to claim 1, wherein the S1 further comprises:

The sensitive information in the self-inspection and reported data is fuzzed by a fuzzy algorithm to obtain fuzzy data, and the self-inspection and reported data is replaced by the fuzzy data to desensitize the self-inspection and reported data.
The method for pushing dynamic epidemic prevention knowledge according to claim 1, wherein the training method for the push model comprises:

Taking the self-examination and reporting data and the knowledge points to be pushed as training samples;

dividing the training sample into a training set and a verification set according to a preset ratio;

Inputting the training set into a preset training model for regression training, and using the verification set every preset period to verify the accuracy of the model; and

When the accuracy rate is greater than the preset threshold, the training is ended, and the push model is obtained.
The push method of dynamic epidemic prevention knowledge as claimed in claim 4, wherein, the push model formula is:

h(t,X(t))=h 0 (t)exp(β 1 X 1 (t)+β 2 X 2 (t)+...β m X m (t))

Among them, h 0 (t) is the reference rate, exp is an exponential function with the base of natural constant e, and the value of constant e is 2.718282;

X(t) represents the feature information set of the t day, X m (t) represents the mth feature information of the t day; β m represents the coefficient corresponding to the mth feature information of the t day;

h(t,X(t)) represents the knowledge point of epidemic prevention with the greatest intervention effect on the characteristic information set input on the t day.
The method for pushing dynamic epidemic prevention knowledge according to claim 1, wherein, before adjusting the format of the feature information included in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, the method further comprises:

Performing discretization processing on the feature information, the discretization processing includes:

Use the dynamic programming method to find all word segmentation combinations of the feature information, calculate the word weights under each word segmentation combination, traverse to obtain the word segmentation combination corresponding to the word weight with the largest value, and obtain the word segmentation combination A discrete text set of feature information under .
The method for pushing dynamic epidemic prevention knowledge according to claim 6, wherein the finding of all word segmentation combinations of the feature information by using a dynamic programming method comprises:

Construct a gradient descent algorithm and an iterative function based on the feature information, solve the segmentation parameters of the iterative function, solve the result value of the gradient descent algorithm according to the difference of the segmentation parameters, and obtain different word segmentation combinations according to the result value .
A push device for dynamic epidemic prevention knowledge, wherein the device includes:

The extraction module is used to obtain the self-examination and reporting data of the target group on the preset date from the data source, and extract the characteristic information of each user in the target group from the self-examination and reporting data;

The analysis module is used to obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, and screen out the updated and unpublished epidemic prevention knowledge to the target population point, and use the selected epidemic prevention knowledge points as the knowledge points to be pushed;

Processing module: used to adjust the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generate an adjusted feature information set; and

Pushing module: used to construct and train a push model according to the self-checking and reporting data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the maximum intervention effect on the feature information set The epidemic prevention knowledge points are pushed to the preset user terminal.
An electronic device, the electronic device includes: a memory and a processor, the memory stores a push program of dynamic epidemic prevention knowledge that can be run on the processor, and the push program of the dynamic epidemic prevention knowledge is stored by the processor The following steps are implemented when executing:

Obtain self-inspection and reporting data of the target group on a preset date from a data source, and extract characteristic information of each user in the target group from the self-inspection and reporting data;

Obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, screen out the updated and unpublished epidemic prevention knowledge points to the target population, and screen The published epidemic prevention knowledge points are used as knowledge points to be pushed;

Adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generating an adjusted feature information set;

Build and train a push model based on the self-checked and reported data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the epidemic prevention knowledge point push with the greatest intervention effect on the feature information set to the default client.
The electronic device of claim 9, wherein the characteristic information includes the user's basic information, going out records, health status, and answers to symptom questions.
The electronic device according to claim 9, wherein the self-inspection report data of the target group on a preset date is obtained from a data source, and characteristic information of each user in the target group is extracted from the self-inspection and report data ,include:

The sensitive information in the self-inspection and reported data is fuzzed by a fuzzy algorithm to obtain fuzzy data, and the self-inspection and reported data is replaced by the fuzzy data to desensitize the self-inspection and reported data.
The electronic device according to claim 9, wherein the training method of the push model comprises:

Taking the self-examination and reporting data and the knowledge points to be pushed as training samples;

dividing the training sample into a training set and a verification set according to a preset ratio;

Inputting the training set into a preset training model for regression training, and using the verification set every preset period to verify the accuracy of the model; and

When the accuracy rate is greater than the preset threshold, the training is ended, and the push model is obtained.
The electronic device of claim 12, wherein the push model formula is:

h(t,X(t))=h 0 (t)exp(β 1 X 1 (t)+β 2 X 2 (t)+...β m X m (t))

Among them, h 0 (t) is the reference rate, exp is an exponential function with the base of natural constant e, and the value of constant e is 2.718282;

X(t) represents the feature information set of the t day, X m (t) represents the mth feature information of the t day; β m represents the coefficient corresponding to the mth feature information of the t day;

h(t,X(t)) represents the knowledge point of epidemic prevention with the greatest intervention effect on the characteristic information set input on the t day.
The electronic device according to claim 9, wherein, before adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, the following steps are also implemented:

Performing discretization processing on the feature information, the discretization processing includes:

Use the dynamic programming method to find all word segmentation combinations of the feature information, calculate the word weights under each word segmentation combination, traverse to obtain the word segmentation combination corresponding to the word weight with the largest value, and obtain the word segmentation combination A discrete text set of feature information under .
The electronic device according to claim 14, wherein the finding all word segmentation combinations of the feature information by using a dynamic programming method comprises:

Construct a gradient descent algorithm and an iterative function based on the feature information, solve the segmentation parameters of the iterative function, solve the result value of the gradient descent algorithm according to the difference of the segmentation parameters, and obtain different word segmentation combinations according to the result value .
A computer-readable storage medium, the computer-readable storage medium includes a push program of dynamic epidemic prevention knowledge, and the push program of the dynamic epidemic prevention knowledge is executed by a processor to achieve the following steps:

Obtain self-inspection and reporting data of the target group on a preset date from a data source, and extract characteristic information of each user in the target group from the self-inspection and reporting data;

Obtain the latest version of the epidemic prevention guide and the previous version of the epidemic prevention guide from the preset database, compare and analyze the two versions of the epidemic prevention guide, screen out the updated and unpublished epidemic prevention knowledge points to the target population, and screen The published epidemic prevention knowledge points are used as knowledge points to be pushed;

Adjusting the format of the feature information contained in the self-inspection and reporting data to a format consistent with the knowledge point to be pushed, and generating an adjusted feature information set;

Build and train a push model based on the self-checked and reported data and the knowledge points to be pushed, input the feature information set into the trained push model, and output the epidemic prevention knowledge point push with the greatest intervention effect on the feature information set to the default client.
17. The computer-readable storage medium of claim 16, wherein the characteristic information includes the user's basic information, outing records, health status, and responses to symptom questions.
The computer-readable storage medium according to claim 16, wherein the self-inspection report data of the target group on a preset date is obtained from a data source, and each user in the target group is extracted from the self-inspection report data characteristic information, including:

The sensitive information in the self-inspection and reported data is fuzzed by a fuzzy algorithm to obtain fuzzy data, and the self-inspection and reported data is replaced by the fuzzy data to desensitize the self-inspection and reported data.
The computer-readable storage medium of claim 16, wherein the training method of the push model comprises:

Taking the self-examination and reporting data and the knowledge points to be pushed as training samples;

dividing the training sample into a training set and a verification set according to a preset ratio;

Inputting the training set into a preset training model for regression training, and using the verification set every preset period to verify the accuracy of the model; and

When the accuracy rate is greater than the preset threshold, the training is ended, and the push model is obtained.
The computer-readable storage medium of claim 19, wherein the push model formula is:

h(t,X(t))=h 0 (t)exp(β 1 X 1 (t)+β 2 X 2 (t)+...β m X m (t))

Among them, h 0 (t) is the reference rate, exp is an exponential function with the base of natural constant e, and the value of constant e is 2.718282;

X(t) represents the feature information set of the t day, X m (t) represents the mth feature information of the t day; β m represents the coefficient corresponding to the mth feature information of the t day;

h(t,X(t)) represents the knowledge point of epidemic prevention with the greatest intervention effect on the characteristic information set input on the t day.