Application of fuzzy complex integral classifier in hospital infection control monitoring system
Through the application of the fuzzy complex integral classifier in the hospital infection control monitoring system, the problem that it is difficult to achieve accurate solution in the existing technology is solved. Using the fuzzy integral theory and the complex domain method, a scientific and objective evaluation of hospital infection management quality indicators was established. The monitoring system optimizes the quality evaluation of hospital infection management.
Patent Information
- Application Number
- CN202410949821.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-16
- Publication Date
- 2024-11-08
AI Technical Summary
The existing hospital infection control monitoring system is difficult to accurately solve classification problems, especially in the learning and search of large-scale complex problems. The existing technology has shortcomings, which makes it impossible to easily achieve effective classification and monitoring.
Using the fuzzy complex collection volume integral classifier, by establishing a joint evaluation fuzzy set of positive and negative evaluations and a fuzzy rule library for final member evaluation on the static data set, the multi-fuzzy complex collection value integral classification model is trained, and the hospital's sensor control monitoring system is used. Indicators are input to the model for classification, and information processing is performed using fuzzy integral theory and complex domain methods.
In the case of insufficient prior knowledge, the advantages of the fuzzy complex value integral classifier were achieved, and the good results of the evaluation of hospital infection management quality indicators were promoted. The intelligent monitoring system for the evaluation of hospital infection management quality indicators based on big data was promoted. The establishment of the hospital's infection management quality evaluation has been optimized.
Smart Images

Figure CN118919034A_ABST
Abstract
Description
Technical Field
[0001] The invention is based on the application of a fuzzy complex collective integral classifier in a hospital infection control monitoring system, belongs to the technical field of hospital infection monitoring, and specifically relates to the technical field of application of a fuzzy complex collective integral classifier in a hospital infection control monitoring system. Background Art
[0002] The prevention and control technologies and methods of hospital infection were created in the mid-19th century. At that time, hospital infection control was only developed and applied at the technical level. By the mid-20th century, organized hospital infection prevention and control work came into being with the development of modern hospital management. With the development of hospital infection control technology and the frequent exchanges of products and technologies brought about by global economic integration, the technical foundation for establishing relevant standards for hospital infection control has been established, and the demand is also increasing.
[0003] Among the 19,500 international standards in 303 categories of the International Organization for Standardization (ISO), there are 6 categories of standards related to hospital infection control, covering hospital infection control-related products, personal protection, air purification, sterilization of medical supplies, etc. The United States is a representative of the commercial health insurance health system. Its hospital infection control-related requirements are mainly issued in the form of standards and guidelines. The standards are issued by the American National Standards Institute, and the guidelines are formulated and issued by the US Centers for Disease Control and Prevention and industry associations related to hospital infection control. The health system of the United Kingdom is a national health care system with a large medical and health government agency. The hospital infection control guidelines of the United Kingdom are mainly formulated and issued by the National Institute for Clinical Quality Management (NICE) established by the British Ministry of Health and the government. The standards of the United Kingdom are formulated by the British Standards Institution (BSI), which has formulated nearly 300 medical standards. The standards related to hospital infection control also focus on the specification of related products. Germany is a social health insurance health system. Its hospital infection control-related documents are divided into three levels. The first level is the law; the second level is the operation manual issued by the Robert Koch Institute; the third level is the technical guidelines issued by the industry association and the standards issued by the German Standards Association.
[0004] In the 1980s, my country began to organize the prevention and control of hospital infections. After nearly 30 years of development, my country's hospital infection control has made great progress. A system including laws, administrative regulations, departmental regulations, normative documents and standards has been initially established, laying a legal foundation for guiding the prevention and control of hospital infections. In 2006, the Professional Committee of Hospital Infection Control Standards of the Ministry of Health was established, and began to formulate hospital infection control standards in an organized and systematic manner. In 2013, the professional committee was renamed the Professional Committee of Hospital Infection Control of the National Health Standards Committee. So far, 38 hospital infection control standards have been established and 8 have been issued. With the increase in the number of standard projects, the framework system of hospital infection control standards in my country has been initially established.
[0005] In 1965, Zadeh proposed the concept of fuzzy sets, marking the birth of fuzzy mathematics. So far, fuzzy mathematics theory has not only permeated with branches in classical mathematics, but also formed many branches of fuzzy mathematics (such as fuzzy topology, fuzzy algebra, etc.), and fuzzy analysis is one of the typical representatives. Moreover, its application has shown great superiority in many fields with fuzzy problems, such as cluster analysis, image recognition, data structure, system evaluation, automatic control, decision-making, optimization, etc. In 1989, Buckley published the paper "Fuzzy complex numbers" in "Fuzzy Sets and Systems", marking the birth of fuzzy complex analysis. Since then, scholars at home and abroad have comprehensively launched a series of studies on fuzzy complex analysis. Buckley.
[0006] The operation is regarded as a mapping. Based on the "expansion principle", the analytical operation of fuzzy complex numbers is given, and the fuzzy complex function and its operation are further discussed.
[0007] In 1993, Zhang Yue conducted a systematic discussion on the concepts of fuzzy complex sets, complex fuzzy sets, fuzzy complex numbers, complex fuzzy numbers, and fuzzy complex functions in his book Fuzzy Random Dynamical Systems. He used the mapping from the complex number field C to [0, 1] to give the concept of fuzzy complex sets, used the regular fuzzy complex sets to give the general concept of fuzzy complex numbers, and combined the cut-set properties of fuzzy sets to give the four arithmetic operations of fuzzy complex numbers. Zhang Yue constructed complex fuzzy sets using real fuzzy numbers, established the basic operations of complex fuzzy sets such as "inclusion, equality, intersection, union, and cut-off", and obtained important conclusions represented by the decomposition theorem of complex fuzzy sets, the representation theorem, and the expansion principle.
[0008] With the emergence of fuzzy sets, fuzzy technology began to be combined with various classification algorithms to deal with cognitive problems that classical classification algorithms cannot handle during the classification process. Fuzzy measures and fuzzy integrals are used as classification tools for classifier design and fusion. Yao Minghai, Wang Xizhao and others first applied fuzzy integrals to classification technology and achieved good results. The main problem of fuzzy integrals in classification technology is the fuzzy measure problem and the fuzzy integral theory given to it. The authors of the literature studied a series of problems such as the measure of complex fuzzy numerical functions, complex fuzzy numerical measurable functions and complex fuzzy numerical integrals on complex fuzzy number sets. These have laid a certain foundation for the application of fuzzy complex analysis theory, proposed the concept of complex fuzzy valued functions, theoretically studied complex fuzzy numerical measurement problems, studied complex fuzzy valued measurable functions and complex fuzzy numerical integrals and other important series of problems, and especially used Sugeno-type fuzzy complex numerical integrals as new tools to apply them to classification fusion technology and initially achieved good results.
[0009] In China, Qu Zhijun and others believe that the construction of big data information technology for hospital infection management based on monitoring data is lagging behind; the research on the fusion of fuzzy complex set-valued integral classifiers and their application in promoting the algorithm of intelligent monitoring system for hospital infection management quality indicators based on big data needs further in-depth research. For this reason, we proposed the application of fuzzy complex set-valued integral classifiers in hospital infection control monitoring systems. Summary of the invention
[0010] The technical problem to be solved by the present invention is to overcome the existing defects and provide an application of a fuzzy complex collective integral classifier in a hospital infection control monitoring system, thereby solving the learning search and precise solution in the classification problem, which has not been easily achieved so far. Many scientific researchers are constantly exploring various heuristic algorithms to approximate solutions to large-scale complex problems.
[0011] To achieve the above object, the present invention provides the following technical solutions:
[0012] The application of fuzzy complex collective integral classifier in hospital infection control monitoring system includes the following steps:
[0013] Step 1: Based on the static data set, establish the fuzzy set of "positive and negative evaluation" joint evaluation;
[0014] Step 2: Establish a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation;
[0015] Step 3: training multi-fuzzy complex set value integral classification model;
[0016] Step 4: Input the indicators of the hospital infection control monitoring system into the fuzzy complex value integral classification model for classification.
[0017] As a preferred technical solution of the present invention, the first step: based on the static data set, establishing a fuzzy set of "positive and negative evaluation" joint evaluation includes the following steps:
[0018] First, describe the information with a language set that has a certain judgment ability, and then digitally mark it, using the method of fuzzy mathematics. Typical indicators described by fuzzy language include "good" and "bad", "more" and "less", "important" and "unimportant", "easy" and "general", expressed as real numbers between 0 and 1. In specific applications, people often use the "scoring" method of industry experts, using real values according to the "degree" of the "meaning" described.
[0019] For description purposes, a is limited to [0.1], where 0 represents (bad, few, unimportant, average), 1 represents (good, many, important, easy), and other meanings.
[0020] According to the cognitive characteristics of the human brain, it is recommended that in practical applications, information should be expressed scientifically and in accordance with human thinking, and the value of information should be converted into a complex number (value) form, recorded as) a+ib(a, b∈[0,1]) a real number in
[0021] Its specific meaning can be seen that the real part a represents the positive evaluation of the evaluation index, and the imaginary part B represents the negative evaluation of the evaluation index. Therefore, the combination of "positive" and "negative" can be used to describe the digitization of information;
[0022] Now assume that we have obtained the index information related to this problem by using the above content. The method is: A = {a1, a2, ...a n}, a k =Rea k +i Im a k ,
[0023] In order to meet the above requirements, the measurement problem of this information is defined:
[0024] X={x1,x2,...x n} is marked as the cause set; Y = {y1, y2, ...y n}Mark as result set;
[0025] σ Y The conditional fuzzy complex value of (· / x) is measured in the ratio X to Y by σ Y (· / x)=Reσ Y (· / x)+i Imσ Y (· / x) indicates;
[0026] g x =Re g x+i Im g x ; g y =Re g y +i Im g y The consideration of representing fuzzy complex-valued measurements on X and Y is improved based on the information obtained. In the case of real values, the information obtained is usually listed as a subset F on Y, which is usually divided into three categories:
[0027] (1) F has only one element, which is deterministic information;
[0028] (2) F has two or more elements, which is non-deterministic information;
[0029] (3) F∈F(Y) This is fuzzy information;
[0030] If the membership function of fuzzy complex-valued information is A, then the measure of fuzzy complex-valued information can be defined as:
[0031]
[0032] include
[0033]
[0034] According to the fuzzy integral theory and key formulas, the relevant formulas of fuzzy real values are extended to complex values:
[0035] g Y (·)=Re g Y (·)+i Im g Y (·)
[0036] Then the information The complex-valued measure of can be expressed as in
[0037] Re g x +i Im g x As a priori fuzzy complex measure, it is used to measure "an element in E is the cause of the result" The ambiguity of this statement:
[0038] The fuzzy complex value is Then blur the information His function is:
[0039]
[0040] Include
[0041]
[0042] As a preferred technical solution of the present invention, the second step: establishing a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation includes the following steps:
[0043] After obtaining the fuzzy information A, the increase of σY(A) is used as a sign that the prior information gX is improved. Now assume that the measures gX(·) and gY(· / x) both satisfy the λ-rule, that is: the real-valued mapping g λ :P(X)→[0,1] satisfies:
[0044] 1) g λ (X) = 1
[0045] 2) like Then we have:
[0046] gλ(A∪B)=gλ(A)+gλ(B)+λgλ(A)·gλ(B),λ>—1
[0047] 3) gλ meets the continuity requirements;
[0048] Obviously, when λ>-1, gλ is a fuzzy real-valued measure,
[0049] Easy to prove:
[0050] (1)
[0051]
[0052] (2)
[0053] (3):
[0055]
[0056] When the domain
[0057] X={x1,x2,...,x n}
[0058] In finite time, we can use the above formula to construct a λ-fuzzy real-valued measure and set
[0059]
[0060] Arrange in descending order, i.e.
[0061] Then we have:
[0062]
[0063] in
[0064] F i ={x1,x2,...,x i}
[0065] By X The monotonicity of (·) shows that there exists an integer l such that
[0066]
[0067] Here l satisfies the inequality:
[0068]
[0069] According to this rule, the whole process of learning must improve Those that contribute to the increase in g X The fuzzy density of the fuzzy matrix is reduced, and the fuzzy density of the fuzzy matrix that does not contribute is reduced.
[0070] Assume g k (1≤k≤n) is g X The fuzzy complex-valued density of
[0071] g k =Reg k +iImReg k , g X =Reg X +iImg X , g Y =Reg Y +i Img Y ,
[0072] Note: Only those g that satisfy 1≤k≤n k Just right Y The value of (A) has an impact on
[0073] Therefore, a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation is established.
[0074] As a preferred technical solution of the present invention, the third step: training a multi-fuzzy complex valued integral classification model comprises the following steps:
[0075] The learning algorithm for the implementation process of constructing the human-like learning model is:
[0076]
[0077]
[0078] Here the real number α is a parameter to adjust the learning speed. The smaller α is, the faster the learning speed will be. i The greater the change, After determining the α value, the fuzzy density obtained after n times of learning according to the above algorithm is the desired ideal learning result;
[0079] After learning, we can get a relatively ideal g k =Reg k +iImReg k , and then get
[0080] g X =Reg X +iImg X , g Y =Reg Y +i Img Y
[0081]
[0082] Compared with the prior art, the beneficial effects of the present invention are: combining the theoretical research of the subject (core data information processing technologies such as data information "classification", "learning", and "measurement") with modern Internet technology and big data technology, and utilizing the fuzzy complex set value integral multi-classifier fusion technology to utilize the performance advantages of each single classifier. In the case of insufficient prior knowledge, the advantages of the fuzzy complex set value integral classifiers of each member can still make the entire classifier system achieve good results. The research on the fusion of fuzzy complex set value integral classifiers can actively promote the evaluation process of hospital infection management quality indicators based on big data, which has important practical significance. Therefore, the fuzzy complex set value integral classifier is studied to explore the problem of hospital infection management and optimize the evaluation of hospital infection management quality indicators. The significance of this project is to establish a hospital infection management quality monitoring system that is in line with the actual norms of the hospital, scientific, and objective.
[0083] The core technical algorithm for solving the problem of hospital infection management quality monitoring data processing is to construct a basic learning criterion algorithm theory based on the fuzzy complex information human-like learning model based on S-type fuzzy complex-valued integral.
[0084] The research on fuzzy complex set-valued integral classifier algorithm model, fuzzy complex valued measure learning algorithm, etc., especially the expansion of the classic fuzzy comprehensive evaluation algorithm into an improved composite fuzzy complex set-valued integral classifier algorithm is original work.
[0085] Secondly, the theoretical research results of data stream classifiers based on fuzzy complex set-valued integrals have certain scientific value and have broad application prospects in the context of big data.
[0086] Thirdly, the algorithm is expanded from the "real value" situation to the "complex value" situation. Relying on the theory of fuzzy complex analysis, the "real part" and "imaginary part" of complex numbers are used to process the "positive" and "negative" comprehensive information of educational information in the evaluation of practical problems in the scientific evaluation of undergraduate vocational education. Fuzzy language variables are used for fuzzy logic reasoning to make the classifier processing technology closer to human brain thinking, thereby increasing the dimension of machine thinking and improving machine intelligence. BRIEF DESCRIPTION OF THE DRAWINGS
[0087] Figure 1 The process of implementing the human-like learning model algorithm of the present invention;
[0088] Figure 2 The S-type fuzzy complex set value integral data stream classifier process of the present invention;
[0089] Figure 3 This is the research idea of the present invention. DETAILED DESCRIPTION
[0090] 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.
[0091] See also Figure 1-3 The present invention provides a technical solution: application of fuzzy complex collective integral classifier in hospital infection control monitoring system, comprising the following steps:
[0092] Step 1: Based on the static data set, establish the fuzzy set of "positive and negative evaluation" joint evaluation;
[0093] Step 2: Establish a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation;
[0094] Step 3: training multi-fuzzy complex set value integral classification model;
[0095] Step 4: Input the indicators of the hospital infection control monitoring system into the fuzzy complex value integral classification model for classification.
[0096] As a preferred technical solution of the present invention, the first step: based on the static data set, establishing a fuzzy set of "positive and negative evaluation" joint evaluation includes the following steps:
[0097] First, describe the information with a language set that has a certain judgment ability, and then digitally mark it, using the method of fuzzy mathematics. Typical indicators described by fuzzy language include "good" and "bad", "more" and "less", "important" and "unimportant", "easy" and "general", expressed as real numbers between 0 and 1. In specific applications, people often use the "scoring" method of industry experts, using real values according to the "degree" of the "meaning" described.
[0098] For description purposes, a is limited to [0.1], where 0 represents (bad, few, unimportant, average), 1 represents (good, many, important, easy), and other meanings.
[0099] According to the cognitive characteristics of the human brain, it is recommended that in practical applications, information should be expressed scientifically and in accordance with human thinking, and the value of information should be converted into a complex number (value) form, recorded as) a+ib(a, b∈[0,1]) a real number in
[0100] Its specific meaning can be seen that the real part a represents the positive evaluation of the evaluation index, and the imaginary part B represents the negative evaluation of the evaluation index. Therefore, the combination of "positive" and "negative" can be used to describe the digitization of information;
[0101] Now assume that we have obtained the index information related to this problem by using the above content. The method is: A = {a1, a2, ...a n}, a k =Rea k +i Im a k ,
[0102] In order to meet the above requirements, the measurement problem of this information is defined:
[0103] X={x1,x2,...x n} is marked as a cause set; Y = {y1, y2, ...y n}Mark as result set;
[0104] σ γ The conditional fuzzy complex value of (· / x) is measured in the ratio X to Y by σ Y (· / x)=Reσ Y (· / x)+i Imσ Y (· / x) indicates;
[0105] g X =Re g X +i Im g X ; g y =Re g y +i Im g yThe consideration of representing fuzzy complex-valued measurements on X and Y is improved based on the information obtained. In the case of real values, the information obtained is usually listed as a subset F on Y, which is usually divided into three categories:
[0106] (1) F has only one element, which is deterministic information;
[0107] (2) F has two or more elements, which is non-deterministic information;
[0108] (3) F∈F(Y) This is fuzzy information;
[0109] If the membership function of fuzzy complex-valued information is A, then the measure of fuzzy complex-valued information can be defined as:
[0110]
[0111] include
[0112]
[0113] According to the fuzzy integral theory and key formulas, the relevant formulas of fuzzy real values are extended to complex values:
[0114] g Y (·)=Re g Y (·)+i Im g Y (·)
[0115] Then the information The complex-valued measure of can be expressed as in
[0116] Re g x +i Im g x As a priori fuzzy complex measure, it is used to measure "an element in E is the cause of the result" The ambiguity of this statement:
[0117] The fuzzy complex value is Then blur the information His function is:
[0118]
[0119] Include
[0120]
[0121] As a preferred technical solution of the present invention, the second step: establishing a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation includes the following steps:
[0122] After obtaining the fuzzy information A, the increase of σY(A) is used as a sign that the prior information gX is improved. Now assume that the measures gX(·) and gY(· / x) both satisfy the λ-rule, that is: the real-valued mapping g λ :P(X)→[0,1] satisfies:
[0123] 2) g λ (X) = 1
[0124] 2) like Then we have:
[0125] gλ(A∪B)=gλ(A)+gλ(B)+λgλ(A)·gλ(B),λ>—1
[0126] 3) gλ meets the continuity requirements;
[0127] Obviously, when λ>-1, gλ is a fuzzy real-valued measure,
[0128] Easy to prove:
[0129] (1)
[0130]
[0131] (2)
[0132] (3):
[0134]
[0135] When the domain
[0136] X={x1,x2,...,x n}
[0137] In finite time, we can use the above formula to construct a λ-fuzzy real-valued measure and set
[0138]
[0139] Arrange in descending order, i.e.
[0140] Then we have:
[0141] n
[0142]
[0143] in
[0144] F i={x1,x2,...,x i}
[0145] By X The monotonicity of (·) shows that there exists an integer l such that
[0146]
[0147] Here l satisfies the inequality:
[0148]
[0149] According to this rule, the whole process of learning must improve Those that contribute to the increase in g X The fuzzy density of the fuzzy matrix is reduced, and the fuzzy density of the fuzzy matrix that does not contribute is reduced.
[0150] Assume g k (1≤k≤n) is g X The fuzzy complex-valued density of
[0151] g k =Reg k +iImReg k , g X =Reg X +iImg X , g Y =Reg Y +i Img Y ,
[0152] Note: Only those g that satisfy 1≤k≤n k Just right Y The value of (A) has an impact on
[0153] Therefore, a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation is established.
[0154] As a preferred technical solution of the present invention, the third step: training a multi-fuzzy complex valued integral classification model comprises the following steps:
[0155] The learning algorithm for the implementation process of constructing the human-like learning model is:
[0156]
[0157] Here the real number α is a parameter to adjust the learning speed. The smaller α is, the faster the learning speed will be. i The greater the change, After determining the α value, the fuzzy density obtained after n times of learning according to the above algorithm is the desired ideal learning result;
[0158] After learning, we can get a relatively ideal gk =Reg k +i Im Reg k , and then get
[0159] g X =Reg X +i Img X , g Y =Reg Y +i Img Y
[0160]
[0161] Design of indicators for hospital infection control monitoring system
[0162] Hospital infection monitoring usually meets the needs of the monitoring subject, and can objectively, accurately and sensitively reflect the risks and practical conditions of hospital infection prevention and control and management, and is described by the data elements contained in the monitoring target system and the results of the monitoring indicators formed by these data elements. In the document "Guidelines for the Implementation of Basic Data Sets and Quality Control Indicator Sets for Hospital Infection Monitoring (2016 Edition)", there are 70 basic data sets for hospital infection monitoring and 64 quality control indicator sets for hospital infection monitoring, which can be considered and evaluated from different dimensions.
[0163] For example, setting 64 hospital infection monitoring quality control indicator sets is a C-type classification problem. The hospital infection monitoring quality control indicator set is composed of X = {x1, x2, ..., X 64} means that the definition of real-valued function on the index set X is f = (f(x1), f(x2), ..., f(X 64 )), the complex part represents the positive part to evaluate the quality control index set of each hospital infection monitoring, and its value is between 0 and 1; 1-f = (1-f (x1), 1-f (x2), ..., 1-f (X 64 )) This represents a negative evaluation, the imaginary part represents a negative subordinate evaluation, and u is set i For i fuzzy complex set valued measure corresponding class, so a k =∫ X fd i , b k =∫ X (1-f)dμ i , the output of the integral classifier is a set of two-dimensional vectors:
[0164]
[0165] (a k ,ib k ) indicates that the example belongs to k positive membership evaluations and negative membership evaluations and the membership evaluation of class d k=a k -ib k , k = 1, 2, ..., 64, the final classification can be given according to the principle of maximum number of members.
[0166] Here, negative membership evaluation is added and the classic integral classifier is improved, and the drawn classifier is more comprehensive and more in line with the form of human thinking.
[0167] Process Analysis of Hospital Infection Control Monitoring System
[0168] In order to simplify our problem analysis, the "Guidelines for the Implementation of Basic Data Sets and Quality Control Indicator Sets for Hospital Infection Monitoring (2016 Edition)" contains 70 basic data sets for hospital infection monitoring and 64 hospital infection monitoring quality control indicator groups, which are simplified into two types of problems with four indicator sets as shown in Table 1 below:
[0169] Table 1 Attribute values of indicator sets
[0170]
[0171] Table 2 shows the fuzzy measure for each category.
[0172]
[0173] according to Figure 1 Algorithm technology process: Get the following Table 3 and Table 4
[0174] Table 3. Positive and negative evaluation results
[0175]
[0176] Table 4 Evaluation results
[0177]
[0178] Take the maximum value in Table 4. The classification results are shown in Table 5.
[0179] Table 5 Classification results
[0180]
[0181] Based on the above improvement of the final membership evaluation using fuzzy reasoning, it can be concluded that the four situations in the above example belong to categories 2, 1, 1 and 1 respectively. The use of fuzzy integral tools, two-way membership evaluation and fuzzy reasoning methods to finally classify the cases reflects the comprehensive classification. The two-way membership evaluation reflects the objectivity of the classification and better achieves the classification effect.
[0182] With the continuous deepening of hospital infection information construction, hospital infection monitoring has obtained a large amount of monitoring data. In the era of big data, how to make good use of massive data to establish a scientific, objective, standardized and hospital infection management quality indicator evaluation system that conforms to the actual situation of the hospital is particularly important for hospital infection management personnel and clinical medical staff, and it also plays a vital role in improving the quality of hospital infection monitoring.
[0183] By establishing an AI-based big data hospital infection management quality control information network system, information technology is used to standardize and dynamically collect the basic information generated by target monitoring objects at all levels in hospital infection management activities, and incorporate it into the national or local hospital infection management quality control basic data set. Quality control entities at all levels that have obtained different management authorizations in the management system use the auxiliary analysis and decision-making functions of the information system to independently or through joint collaboration conduct targeted big data artificial intelligence analysis on the target data information collected and obtained by the information system, and determine response strategies.
[0184] 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. Application of fuzzy complex collective integral classifier in hospital infection control monitoring system, characterized by: The following steps are involved: Step 1: Based on the static data set, establish the fuzzy set of "positive and negative evaluation" joint evaluation; Step 2: Establish the fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation; Step 3: training multi-fuzzy complex set value integral classification model; Step 4: Input the indicators of the hospital infection control monitoring system into the fuzzy complex value integral classification model for classification.
2. The application of the fuzzy complex collective integral classifier in a hospital infection control monitoring system according to claim 1 is characterized in that: The first step: on the basis of the static data set, establishing the fuzzy set of "positive and negative evaluation" joint evaluation includes the following steps: First, describe the information with a language set that has a certain judgment ability, and then digitally mark it, using the method of fuzzy mathematics. The indicators described by typical fuzzy language include "good" and "bad", "more" and "less", "important" and "unimportant", "easy" and "general", and are represented by real numbers between 0 and 1. In specific applications, people often use the "scoring" method of industry experts, using real values according to the "degree" of the "meaning" described. For description purposes, a is limited to [0.1], where 0 represents (bad, few, unimportant, average), 1 represents (good, many, important, easy), and other meanings. According to the cognitive characteristics of the human brain, it is recommended that in practical applications, information should be expressed scientifically and in accordance with human thinking, and the value of information should be converted into a complex number (value) form, recorded as) a+ib(a,b∈[0,1]) a real number in Its specific meaning can be seen that the real part a represents the positive evaluation of the evaluation index, and the imaginary part B represents the negative evaluation of the evaluation index. Therefore, the combination of "positive" and "negative" can be used to describe the digitization of information; Now assume that we have obtained the index information related to this problem by using the above content. The method is: A = {a1, a2, ...a m },a k =Re a k +i Im a k , In order to meet the above requirements, the measurement problem of this information is defined: X={x1,x2,...x n } is marked as a cause set; Y = {y1, y2, ...y n }Mark as result set; σ γ The conditional fuzzy complex value of (· / x) is measured in the ratio X to Y by σ Y (· / x)=Reσ Y (· / x)+iIMσ Y (· / x) indicates; g x =Reg x +iImgg x ; g y =Reg y +iImg y The consideration of representing fuzzy complex-valued measurements on X and Y is improved based on the information obtained. In the case of real values, the information obtained is usually listed as a subset F on Y, which is usually divided into three categories: (1) F has only one element, which is deterministic information; (2) F has two or more elements, which is non-deterministic information; (3) F∈F(Y) This is fuzzy information; If the membership function of fuzzy complex-valued information is A, then the measure of fuzzy complex-valued information can be defined as: include According to the fuzzy integral theory and key formulas, the relevant formulas of fuzzy real values are extended to complex values: g Y (·)=Reg Y (·)+iImg Y (·) Then the information The complex-valued measure of can be expressed as in Reg x +iImg x As a priori fuzzy complex measure, it is used to measure "an element in E is the cause of the result" The ambiguity of this statement: The fuzzy complex value is Then blur the information His function is: Include 3. The application of the fuzzy complex collective integral classifier in a hospital infection control monitoring system according to claim 2 is characterized in that: The second step: establishing a fuzzy rule base for the joint evaluation of "positive and negative evaluation" and the final member evaluation includes the following steps: After obtaining the fuzzy information A, the increase of σY(A) is used as a sign that the prior information gX is improved. Now assume that the measures gX(·) and gY(· / x) both satisfy the λ-rule, that is: the real-valued mapping g λ :P(X)→[0,1] satisfies: 1)g λ (X)=1 2) like Then we have: gλ(A∪B)=gλ(A)+gλ(B)+λgλ(A)·gλ(B),λ>—1 3) gλ meets the continuity requirements; Obviously, when λ>-1, gλ is a fuzzy real-valued measure, Easy to prove: (1) (2) (3): When the domain X={x1,x2,...,x n } In finite time, we can use the above formula to construct a λ-fuzzy real-valued measure and set Arrange in descending order, i.e. Then we have: in F i ={x1,x2,...,x i } By X The monotonicity of (·) shows that there exists an integer l such that Here l satisfies the inequality: According to this rule, the whole process of learning must improve Those that contribute to the increase in g X The blur density of the fuzzy matrix is reduced, and the blur density of the fuzzy matrix that does not contribute is reduced. Assume g k (1≤k≤n) is g X The fuzzy complex-valued density of g k =Reg k +iImReg k ,g X =Reg X +iImg X ,g Y =Reg Y +iImg Y , Note: Only those g that satisfy 1≤k≤n k Just right Y The value of (A) has an impact, Therefore, a fuzzy rule base for "positive and negative evaluation" joint evaluation and final member evaluation is established.
4. The application of the fuzzy complex collective integral classifier in a hospital infection control monitoring system according to claim 3 is characterized in that: The third step: training the multi-fuzzy complex set-valued integral classification model comprises the following steps: The learning algorithm for the implementation process of constructing the human-like learning model is: Here the real number α is a parameter to adjust the learning speed. The smaller α is, the faster the learning speed will be. i The greater the change, After determining the α value, the fuzzy density obtained after n times of learning according to the above algorithm is the desired ideal learning result; After learning, we can get a relatively ideal g k =Reg k +iImReg k , and then get g X =Reg X +iImg X ,g Y =Reg Y +iImg Y
