A decision-making method for AIDS prevention and treatment based on multi-data analysis model

Through multi-data analysis model and decision tree algorithm, the problem of slow progress in AIDS prevention and control caused by manual calculations is solved, automatic calculation and data analysis of AIDS-infected patients and patients management is realized, effective control of infection sources, and the efficiency of AIDS prevention and control work is improved .

CN112635074BActive Publication Date: 2024-11-22YUNNAN PROVINCIAL CENT FOR DISEASE CONTROL & PREVENTION
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202011515344.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-12-21
Publication Date
2024-11-22
Estimated Expiration
2040-12-21

AI Technical Summary

Technical Problem

The existing AIDS prevention and control decision-making methods based on multi-data analysis models rely on manual calculations, resulting in a long time and heavy burden on the grassroots follow-up persons, and the inability to analyze data in the jurisdiction in a timely manner, resulting in a slow progress in AIDS prevention and control work.

Method used

By establishing a basic data collection model, using statistical description statistical calculation indicators, analytical and statistical models of case epidemiological investigation, PITC, treatment referral and three types of population ART progress data are constructed, and a decision tree algorithm is used to generate auxiliary Decision-making measures, including solidified indicator data warning, abnormal case overcurrent rate warning, etc., to realize automatic calculation and data analysis.

Benefits of technology

It has realized automatic calculation and data analysis of comprehensive management of AIDS-infected patients and patients, provided timely data analysis services, effectively controlled the source of infection, reduced new infections, and improved the efficiency and progress of AIDS prevention and control work.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112635074B_ABST
    Figure CN112635074B_ABST
Patent Text Reader

Abstract

The invention discloses an AIDS prevention and treatment decision-making method based on a multi-data analysis model. The specific steps of the prevention and treatment decision-making method are as follows: step one: establishing a basic data collection model; step two: applying descriptive statistics in statistics to the basic data collected in step one, and constructing an analysis statistical model according to corresponding indicators to obtain a descriptive chart; step three: monitoring the descriptive chart in step two, selecting feature information of attributes of different categories contained in the descriptive chart, and forming a feature classification set; step four: constructing a decision tree for the feature classification set using a decision tree algorithm, and recursively generating child nodes from top to bottom according to feature evaluation criteria to obtain auxiliary decision-making measures; the prevention and treatment decision-making method provides an auxiliary basis for AIDS prevention and treatment work, and at the same time carries out targeted follow-up on AIDS prevention and treatment, which is conducive to better achieving the goal of AIDS prevention and treatment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of comprehensive management of HIV-infected persons and patients, and in particular to an AIDS prevention and treatment decision-making method based on a multi-data analysis model. Background Art

[0002] Data analysis is an important part of the comprehensive management of HIV-infected people and patients. At present, most people responsible for AIDS follow-up have to manually calculate the progress of the comprehensive management of infected people and patients every month, and compile databases such as the list of untested CD4, untested spouses, and unchecked tuberculosis to guide the work. This work takes a long time. At the same time, the grassroots follow-up responsible persons have a lot of work to do and cannot analyze the relevant data in their jurisdiction in a timely manner, resulting in a slow progress in AIDS prevention and control work.

[0003] The existing AIDS prevention and control decision-making method based on multiple data analysis models uses manual calculations for the comprehensive management progress of infected persons and patients. This method takes a long time to work. At the same time, the grassroots follow-up responsible persons have a lot of work to do and cannot analyze the relevant data in the jurisdiction in a timely manner, resulting in a slow progress in AIDS prevention and control work. For this reason, we propose an AIDS prevention and control decision-making method based on multiple data analysis models. Summary of the invention

[0004] The purpose of the present invention is to provide an AIDS prevention and control decision-making method based on a multi-data analysis model to solve the problem that the existing AIDS prevention and control decision-making method based on a multi-data analysis model proposed in the above background technology adopts manual calculation for the comprehensive management progress of infected persons and patients. This method takes a long time to work, and the grassroots follow-up responsible persons have a lot of work to do and cannot analyze the relevant data in the jurisdiction in time, resulting in slow progress in AIDS prevention and control work.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solution: an AIDS prevention and treatment decision-making method based on a multi-data analysis model, wherein the specific steps of the prevention and treatment decision-making method are as follows:

[0006] Step 1: Establish a basic data collection model to collect basic data of AIDS patients;

[0007] Step 2: Apply descriptive statistics in statistics to the basic data collected in step 1, calculate the corresponding indicators, and build analytical statistical models for case flow survey, PITC, treatment referral and ART progress data of the three groups of people based on the corresponding indicators, and obtain descriptive charts of patient data, case flow survey data, PITC information, treatment referral information and ART progress data of the three groups of people;

[0008] Step 3: monitor the descriptive chart in step 2, select feature information of attributes of different categories contained in the descriptive chart, and form a feature classification set;

[0009] Step 4: Use the decision tree algorithm to construct a decision tree for the feature classification set, and recursively generate child nodes from top to bottom according to the feature evaluation criteria until the data set is indivisible, then stop the decision tree from growing and obtain auxiliary decision measures.

[0010] Preferably, the basic data include personal information, home address, route of infection, whether imprisoned and whether alive; the descriptive charts include histograms, line graphs and pie charts; the corresponding indicators include average, arithmetic sum, mean, mode, variance, standard deviation, maximum and minimum values; the characteristic information includes CD4 detection, case flow survey, treatment referral, antiviral treatment and PITC AIDS testing consultation services; the auxiliary decision-making measures include solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive conversion reminder, death reminder, detention release reminder and other indicator warnings and message push; the decision tree is specifically a decision tree between single model data or multiple model data.

[0011] Preferably, the specific steps of calculating the corresponding indicators by applying descriptive statistics in statistics in step 2 are as follows:

[0012] (1) Mean calculation: Different classification indicators are statistically calculated through mean calculation. The mean calculation formula is as follows:

[0013]

[0014] It can be concluded that: in the current environment, the observed values ​​of each classification index are relatively concentrated in the central position. The mean value can be used to observe the treatment situation in each region and further analyze the reasons for low indicators;

[0015] (2) Calculation of variance and standard deviation: Based on the mean formula in (1), calculate the variance and the degree of deviation of the statistical indicator data. The calculation formulas for variance and standard deviation are as follows:

[0016] variance:

[0017] Standard Deviation:

[0018] It can be concluded that the smaller the variance s2 is, the smaller the deviation of the indicator data is; this indicates that the CD4 monitoring rate in the region fluctuates greatly, and the reason may be the high HIV transmission rate; at the same time, the calculated variance is used to calculate the standard deviation, and the standard deviation is used to determine the coordinate interval of the histogram and line graph, making the data display diagram clearer;

[0019] (3) Mode search: refers to the value with a clear central trend point in the statistical distribution, that is, the value with the most occurrences; this method uses a binary search tree algorithm to find the data with the highest frequency. The steps are as follows:

[0020] ①Traverse the binary tree and use a hash table to count;

[0021] ② The sequence after in-order traversal is an ordered sequence, and the majority is found during in-order traversal;

[0022] (4) Maximum and minimum value search:

[0023] Through the sequential search algorithm, the maximum and minimum values ​​in different regions or different time periods are found, and the work status indicator data in the regional or time period dimension is analyzed.

[0024] Preferably, the feature classification set includes converting case epidemiological survey into confirmed, infected, and normal diagnostic information feature classification sets, dividing case epidemiological survey rate intervals, forming case epidemiological survey rate feature classification sets according to different interval values, forming referral, dropout, and enrollment information and dropout, non-referral, and non-enrollment feature classification sets according to referral information, and forming treated and untreated feature classification sets according to ART progress.

[0025] Preferably, the decision tree algorithm idea is as follows:

[0026] S1: Starting from the root node, each feature of the training data is calculated, and the training data is distributed to its child nodes (branches) according to the uncertainty of each feature. Along the branch, it may reach a leaf node or another internal node;

[0027] S2: Recursively execute the remaining features until a leaf node is reached; when all leaf nodes are reached, a decision tree is obtained.

[0028] Preferably, the decision tree uses category entropy (H(c)) and feature conditional entropy (H(c|x)), and on this basis, calculates information gain (G(x)). In the process of establishing a decision tree, the larger the information gain of a feature, the stronger the ability of the feature to reduce the entropy of the sample, and the stronger the ability of this feature to change data from uncertainty to certainty. In the case of multiple categories, the information gains of different categories are compared, and the category with the largest information gain value is taken as the root node for splitting to generate a decision tree. The formula is as follows:

[0029]

[0030] G(x)=H(c)-H(c|x) (2).

[0031] Preferably, the generation of the decision tree comprises the following steps:

[0032] SS1: Feature selection: Feature selection refers to selecting a feature from the features of the training data as the splitting point of the current node;

[0033] SS2: Training data set: This data set is a set formed by a feature in different feature classification sets and a feature in another one or more feature classification sets, wherein each data set formed corresponds to a decision basis, and the formed decision basis is used to remind patients and medical staff of untreated conditions; the decision basis formed according to the training sample data set is used to provide a template for learning the decision tree;

[0034] SS3: Generate decision tree: According to the selected feature evaluation criteria, recursively generate child nodes from top to bottom, and further improve the decision tree. In the same way, calculate the remaining three dimensions until the data set is indivisible, then stop growing the decision tree.

[0035] Preferably, the split point includes a feature classification set of treatment and untreatment formed by ART progression; and the decision basis includes the confirmed features in the PITC diagnostic information feature classification set + the enrollment features in the referral information feature set + the untreated in the ART progression features.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows: the AIDS prevention and control decision-making method based on the multi-data analysis model proposes an auxiliary decision-making method for the work required for AIDS prevention and control, such as automatic calculation of comprehensive management township data, statistical analysis, analysis and statistics of case flow survey data, calculation and analysis and statistics of PITC data, management and statistical analysis of antiviral treatment referral information, calculation and statistical analysis of three types of population data, data quality control, other reports and timed reminder information push, etc., wherein the auxiliary decision-making measures include solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive conversion reminder, death reminder, detention release reminder and other indicator warnings and message push, thereby realizing the automatic calculation of indicators related to the comprehensive management of AIDS infected persons and patients, which is conducive to timely providing data analysis services for follow-up responsible persons, and effectively controlling the source of infection and reducing new infections;

[0037] The AIDS prevention and control decision-making method based on multiple data analysis models consists of a basic data collection model, a case flow survey data statistical analysis model established by descriptive statistics in applied statistics, a PITC information statistical analysis model, a treatment referral information statistical analysis model, a statistical analysis model for ART progress data of three groups of people, and an information early warning and reminder method for AIDS patients and medical workers based on a decision tree. The multiple data analysis model is used to obtain basic information, referral information, antiviral treatment information of HIV-infected persons and patients, and information on AIDS testing and consultation services proactively provided by medical staff to patients during medical services. A decision-making algorithm is used to provide a basis for auxiliary decision-making for AIDS prevention and control work, which is conducive to targeted follow-up of AIDS prevention and control work and to better achieve the overall goal of AIDS prevention and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] Figure 1 It is a schematic diagram of the overall structure of the present invention;

[0039] Figure 2 is a flow chart of the decision tree algorithm of the present invention;

[0040] Figure 3 It is a schematic diagram of the structure of the decision tree of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1-Figure 3 The present invention provides a technical solution: an AIDS prevention and treatment decision-making method based on a multi-data analysis model, and the specific steps of the prevention and treatment decision-making method are as follows:

[0043] Step 1: Establish a basic data collection model to collect basic data of AIDS patients;

[0044] Step 2: Apply descriptive statistics in statistics to the basic data collected in step 1, calculate the corresponding indicators, and build analytical statistical models for case flow survey, PITC, treatment referral and ART progress data of the three groups of people based on the corresponding indicators, and obtain descriptive charts of patient data, case flow survey data, PITC information, treatment referral information and ART progress data of the three groups of people;

[0045] Step 3: monitor the descriptive chart in step 2, select feature information of attributes of different categories contained in the descriptive chart, and form a feature classification set;

[0046] Step 4: Use the decision tree algorithm to construct a decision tree for the feature classification set, and recursively generate child nodes from top to bottom according to the feature evaluation criteria until the data set is indivisible, then stop the decision tree from growing and obtain auxiliary decision measures.

[0047] Basic data include personal information, home address, route of infection, whether imprisoned and whether alive; descriptive charts include histograms, line graphs and pie charts; corresponding indicators include average, arithmetic sum, mean, mode, variance, standard deviation, maximum and minimum values; characteristic information includes CD4 testing, case flow survey, treatment referral, antiviral treatment and PITC AIDS testing consultation services; auxiliary decision-making measures include solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive conversion reminder, death reminder, detention release reminder and other indicator warnings and message push; the decision tree is specifically a decision tree between single model data or multiple model data.

[0048] Step 2: Apply descriptive statistics in statistics to calculate the corresponding indicators as follows:

[0049] (1) Mean calculation: Different classification indicators are statistically calculated through mean calculation. The mean calculation formula is as follows:

[0050]

[0051] It can be concluded that: in the current environment, the observed values ​​of each classification index are relatively concentrated in the central position. The mean value can be used to observe the treatment situation in each region and further analyze the reasons for low indicators;

[0052] (2) Calculation of variance and standard deviation: Based on the mean formula in (1), calculate the variance and the degree of deviation of the statistical indicator data. The calculation formulas for variance and standard deviation are as follows:

[0053] variance:

[0054] Standard Deviation:

[0055] It can be concluded that the smaller the variance s2 is, the smaller the deviation of the indicator data is; this indicates that the CD4 monitoring rate in the region fluctuates greatly, and the reason may be the high HIV transmission rate; at the same time, the calculated variance is used to calculate the standard deviation, and the standard deviation is used to determine the coordinate interval of the histogram and line graph, making the data display diagram clearer;

[0056] (3) Mode search: refers to the value with a clear central trend point in the statistical distribution, that is, the value with the most occurrences; this method uses a binary search tree algorithm to find the data with the highest frequency. The steps are as follows:

[0057] ①Traverse the binary tree and use a hash table to count;

[0058] ② The sequence after in-order traversal is an ordered sequence, and the majority is found during in-order traversal;

[0059] (4) Maximum and minimum value search:

[0060] Through the sequential search algorithm, the maximum and minimum values ​​in different regions or different time periods are found, and the work status indicator data in the regional or time period dimension is analyzed.

[0061] The feature classification set includes converting case epidemiological investigation into confirmed, infected, and normal diagnostic information feature classification sets, dividing case epidemiological investigation rate intervals, forming case epidemiological investigation rate feature classification sets according to different interval values, forming referral, dropout, and enrollment information according to referral information, and feature classification sets of dropout, non-referral, and non-enrollment, and forming treated and untreated feature classification sets according to ART progress.

[0062] The idea of ​​the decision tree algorithm is as follows:

[0063] S1: Starting from the root node, each feature of the training data is calculated, and the training data is distributed to its child nodes (branches) according to the uncertainty of each feature. Along the branch, it may reach a leaf node or another internal node;

[0064] S2: Recursively execute the remaining features until a leaf node is reached; when all leaf nodes are reached, a decision tree is obtained.

[0065] The decision tree uses category entropy (H(c)) and feature conditional entropy (H(c|x)), and calculates information gain (G(x)) on this basis. In the process of building a decision tree, the larger the information gain of a feature, the stronger the ability of the feature to reduce the entropy of the sample, and the stronger the ability of this feature to turn uncertainty into certainty of data; in the case of multiple categories, compare the information gains of different categories, take the category with the largest information gain value as the root node for splitting, and generate a decision tree; the formula is as follows:

[0066]

[0067] G(x)=H(c)-H(c|x) (2).

[0068] The generation of a decision tree consists of the following steps:

[0069] SS1: Feature selection: Feature selection refers to selecting a feature from the features of the training data as the splitting point of the current node;

[0070] SS2: Training data set: This data set is a set formed by a feature in different feature classification sets and a feature in another one or more feature classification sets, wherein each data set formed corresponds to a decision basis, and the formed decision basis is used to remind patients and medical staff of untreated conditions; the decision basis formed according to the training sample data set is used to provide a template for learning the decision tree;

[0071] SS3: Generate decision tree: According to the selected feature evaluation criteria, recursively generate child nodes from top to bottom, and further improve the decision tree. In the same way, calculate the remaining three dimensions until the data set is indivisible, then stop growing the decision tree.

[0072] The split point includes the feature classification set of treatment and untreatment formed by ART progression; the decision basis includes the diagnosis features in the PITC diagnosis information feature classification set + the enrollment features in the referral information feature set + the untreated in the ART progression features.

[0073] Working principle: For this type of AIDS prevention and control decision-making method based on multiple data analysis models, the first step is to establish a basic data collection model, and use the basic data collection model to collect basic data of AIDS patients; the second step is to apply the calculation method of descriptive statistics to the collected basic data to construct a variety of data analysis statistical models, including case flow survey data statistical analysis model, PITC information statistical analysis model, treatment referral information statistical analysis model and ART progress data statistical analysis model for three groups of people, to obtain descriptive charts, which include CD4 detection, case flow survey, treatment referral, antiviral treatment and PITC AIDS testing consultation service feature information of AIDS infected persons and patients; the third step is to monitor the descriptive charts in the second step, select the feature information of different categories of attributes contained in the descriptive charts, and form a feature classification set; the fourth step is to use the decision tree algorithm to construct a decision tree for the feature classification set, and recursively generate from top to bottom according to the selected feature evaluation criteria. The data set is divided into child nodes until the decision tree stops growing and auxiliary decision measures are obtained, wherein the auxiliary decision measures include solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive transfer reminder, death reminder, detention release reminder and other indicator warnings and message push; the present invention uses the existing feature classification set, according to the solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive transfer reminder, death reminder, detention release reminder and other indicator warnings and message push and different decision-making basis and feature evaluation criteria, extracts sample data, trains data sets, and constructs different decision trees. The decision basis generated by the decision tree provides data analysis services for the responsible person, provides auxiliary decision-making basis for AIDS prevention and control, and at the same time, carries out targeted follow-up on AIDS prevention and control work, so as to better achieve the goal of AIDS prevention and control.

[0074] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A decision-making method for AIDS prevention and treatment based on a multi-data analysis model, characterized in that; The specific steps of the prevention and control decision-making method are as follows: Step 1: Establish a basic data collection model to collect basic data of AIDS patients; Step 2: Apply descriptive statistics in statistics to the basic data collected in step 1, calculate the corresponding indicators, and build analytical statistical models for case flow survey, PITC, treatment referral and ART progress data of the three groups of people based on the corresponding indicators, and obtain descriptive charts of patient data, case flow survey data, PITC information, treatment referral information and ART progress data of the three groups of people; Step 3: monitor the descriptive chart in step 2, select feature information of attributes of different categories contained in the descriptive chart, and form a feature classification set; Step 4: Use the decision tree algorithm to construct a decision tree for the feature classification set, and recursively generate child nodes from top to bottom according to the feature evaluation criteria until the data set is indivisible, then stop the decision tree from growing, and obtain auxiliary decision measures; The specific steps for calculating the corresponding indicators in the descriptive statistics of applied statistics are as follows: (1) Mean calculation: Different classification indicators are statistically calculated through mean calculation. The mean calculation formula is as follows: It can be concluded that: in the current environment, the observed values ​​of each classification index are relatively concentrated in the central position. The mean value can be used to observe the treatment situation in each region and further analyze the reasons for low indicators; (2) Calculation of variance and standard deviation: Based on the mean formula in (1), calculate the variance and the degree of deviation of the statistical indicator data. The calculation formulas for variance and standard deviation are as follows: variance: Standard Deviation: It can be concluded that the smaller the variance s2 is, the smaller the deviation of the indicator data is; this indicates that the CD4 monitoring rate in the region fluctuates greatly, and the reason may be the high HIV transmission rate; at the same time, the calculated variance is used to calculate the standard deviation, and the standard deviation is used to determine the coordinate interval of the histogram and line graph, making the data display diagram clearer; (3) Mode search: refers to the value with a clear central trend point in the statistical distribution, that is, the value with the most occurrences; this method uses a binary search tree algorithm to find the data with the highest frequency. The steps are as follows: ①Traverse the binary tree and use a hash table to count; ② The sequence after in-order traversal is an ordered sequence, and the majority is found during in-order traversal; (4) Maximum and minimum value search: Through the sequential search algorithm, the maximum and minimum values ​​in different regions or different time periods are found, and the working status indicator data in the regional or time period dimension is analyzed; The idea of ​​the decision tree algorithm is as follows: S1: Starting from the root node, each feature of the training data is calculated, and the training data is distributed to its child nodes according to the uncertainty of each feature. The child nodes may reach the leaf node or another internal node. S2: Recursively execute the remaining features until a leaf node is reached; when all leaf nodes are reached, a decision tree is obtained; The generation of the decision tree comprises the following steps: SS1: Feature selection: Feature selection refers to selecting a feature from the features of the training data as the splitting point of the current node; SS2: Training data set: This data set is a set formed by a feature in different feature classification sets and a feature in another one or more feature classification sets, wherein each data set formed corresponds to a decision basis, and the formed decision basis is used to remind patients and medical staff of untreated conditions; the decision basis formed according to the training sample data set is used to provide a template for learning the decision tree; SS3: Generate decision tree: According to the selected feature evaluation criteria, recursively generate child nodes from top to bottom, and further improve the decision tree. In the same way, calculate the remaining three dimensions until the data set is indivisible, then stop growing the decision tree. The split point includes a feature classification set of treatment and untreatment formed by ART progression; the decision basis includes the confirmed features in the PITC diagnostic information feature classification set + the enrollment features in the referral information feature set + the untreated in the ART progression features.

2. The AIDS prevention and treatment decision-making method based on a multi-data analysis model according to claim 1, characterized in that: The basic data include personal information, home address, route of infection, whether incarcerated and whether alive; the descriptive charts include histograms, line graphs and pie charts; the corresponding indicators include average value, arithmetic sum, mean, mode, variance, standard deviation, maximum value and minimum value; The characteristic information includes CD4 testing, case flow survey, treatment referral, antiviral treatment and PITC AIDS testing consultation services; the auxiliary decision-making measures include solidified indicator data warning, abnormal case flow survey rate warning, abnormal PITC information warning, abnormal treatment referral information warning, treatment reminder, follow-up reminder, spouse positive conversion reminder, death reminder, detention release reminder and other indicator warnings and message push; the decision tree is specifically a decision tree between single model data or multiple model data.

3. The AIDS prevention and treatment decision-making method based on a multi-data analysis model according to claim 2, characterized in that: The feature classification set includes forming a diagnosis information feature classification set of confirmed, infected, and normal cases through epidemiological investigation, dividing the case epidemiological investigation rate intervals, forming a case epidemiological investigation rate feature classification set according to different interval values, forming a referral, dropout, and enrollment information and a feature classification set of dropout, non-referral, and non-enrollment according to referral information, and forming a treated and untreated feature classification set according to the progress of ART.

4. The AIDS prevention and treatment decision-making method based on a multi-data analysis model according to claim 1, characterized in that: The decision tree uses the category entropy H(c) and the feature condition entropy H(c|x), and calculates the information gain G(x) on this basis. In the process of establishing the decision tree, the larger the information gain of a feature, the stronger the ability of the feature to reduce the entropy of the sample, and the stronger the ability of this feature to turn uncertainty into certainty of data; in the case of multiple categories, the information gains of different categories are compared, and the category with the largest information gain value is taken as the root node for splitting to generate a decision tree; The formula is as follows: G(x)=H(c)-H(c|x) (2) In the formula, P(c i ) represents the probability of the i-th category appearing; n represents the number of categories.