Characteristic fuzzy model construction method for infectious disease patients

Abstract

The invention discloses a feature fuzzy model construction method for infectious disease patients, which comprises the following steps of: 1, constructing a comprehensive patient feature set by collecting medical records, examination reports, medical histories and living habit information of the patients; 2, calculating representativeness, uniqueness and noise immunity scores of each patient, and then selecting patient representatives with characterization capabilities; 3, a fuzzy interval type-2 granularity model containing fuzzy feature descriptions of various patients is constructed around the high total score representation; according to the method, the patient group model which is better in anti-noise capability, higher in interpretability and more in information content can be quickly constructed under the condition that only a small amount of disease related data exists, so that modeling can be quickly performed on the susceptible crowd, fuzzy characteristics influencing disease infection are described, and doctors are assisted in knowing infectious disease characteristics and susceptible crowd characteristics.

Description

Technical Field

[0001] The invention belongs to the field of data mining, and in particular is an infection population characteristic modeling algorithm oriented to patient data. Background Art

[0002] With the rapid development of big data technology, the use of data mining technology to assist doctors in understanding disease characteristics and designing personalized treatment plans has become a hot innovation area. Medical patient data refers to data related to disease prevention and emergency response collected during the process of disease prevention and health management. However, big data usually has a large amount of data and contains inevitable errors in the statistical process. It is difficult for the human brain alone to extract valuable information from it.

[0003] Clustering technology is one of the most important tasks in unsupervised learning. Its purpose is to divide data samples into different clusters so that data points in the same group are more similar than data points in different groups. Clustering algorithms usually generate multiple clusters and data prototypes of the clusters. This feature makes it often used to pre-group data and use data prototypes to characterize cluster information. However, the clusters generated by traditional clustering technology have the following problems, which prevents them from providing doctors with more valuable information.

[0004] 1) Contains little information. The clusters formed by the existing clustering algorithms contain too little information. The description of the cluster only contains the patient feature information representing the cluster, and there is no information such as the feature range and feature membership of the cluster.

[0005] 2) Poor interpretability. The clusters formed by existing clustering algorithms are either completely irregular shapes or ordinary circles. These shapes cannot effectively represent the range of valuable features, making it difficult for users to understand the results.

[0006] 3) Poor noise resistance. The algorithm does not consider eliminating the noise in the data set, so that the final cluster model is disturbed by the noise data when the data volume is large. Summary of the invention

[0007] In order to solve the shortcomings of the above-mentioned current technologies, the present invention proposes a method for constructing a feature fuzzy model for patients with infectious diseases, so that a patient group model with better noise resistance, stronger interpretability and more information can be quickly constructed with only a small amount of disease-related data, thereby quickly modeling the susceptible population, describing the fuzzy features that affect disease infection, and assisting doctors in understanding the characteristics of infectious diseases and susceptible populations.

[0008] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical solutions:

[0009] The characteristic fuzzy model construction method for infectious disease patients of the present invention is characterized in that it comprises the following steps:

[0010] Step 1: Collect various characteristics of infectious disease patients and build an infectious disease patient dataset , Indicates Infectious disease patients The jth feature in is the total number of patients with infectious diseases, is the total number of characteristics of each infectious disease patient;

[0011] Step 2: Select representatives of infectious disease patients;

[0012] Step 2.1: Use DBSCAN algorithm to Divide into Infectious disease clusters , and count the number of infectious disease patients in each infectious disease patient cluster and the number of patients with noise-related infectious diseases ,in, Indicates Clusters of infectious patients, express The number of patients with infectious diseases ,lie in Any qth infectious disease patient in , ;

[0013] Step 2.2: Calculate the number of infectious disease patient representatives ,in, is the selected ratio, so according to formula (3) Infectious disease clusters The number of infectious disease patient representatives in ;

[0014] (3)

[0015] Step 2.3, calculation After sorting the Euclidean distance between any two infectious disease patients in descending order, select the first The mean of the Euclidean distances is The density radius , thus obtaining The density radius ;

[0016] Step 2.4: Calculate using formula (5) The qth infectious disease patient Representative score of ;

[0017] (5)

[0018] In formula (5), Represents a positive activation function; express The qth infectious patient With the hth infectious disease patient The Euclidean distance between

[0019] Step 2.5: Calculate according to formula (6) The qth infectious disease patient Uniqueness score After normalization, the normalized patient uniqueness score is obtained ;

[0020] (6)

[0021] In formula (6), express The hth infectious disease patient Representative scores of

[0022] Step 2.6: Calculate according to formula (8) The qth infectious disease patient The non-noise score :

[0023] (8)

[0024] In formula (8), express Middle Infectious disease patients No. Features, express Middle The minimum value of the features, express Middle The maximum value of features;

[0025] Step 2.7: Calculate according to formula (9) Middle Infectious disease patients Final candidate score :

[0026] (9)

[0027] Step 2.8: The final candidate scores of all infectious disease patients in the Infectious disease patients Representatives of infectious disease patients , thus obtaining Infectious disease clusters Representatives of infectious disease patients ,in, express The sth infectious disease patient representative;

[0028] Step 3: Based on the infectious disease patient representatives selected in step 2, a characteristic fuzzy model of the infectious disease patient population is constructed;

[0029] Step 3.1: Construct the left endpoint fitness value of the feature fuzzy model and use the genetic algorithm to Processing is performed to obtain the optimal left endpoint of the characteristic fuzzy model of the infectious disease patient population ,in, express Around The optimal left endpoint of the constructed s-th feature fuzzy model on the j-th feature;

[0030] Step 3.2: Construct the right endpoint fitness value of the feature fuzzy model and use the genetic algorithm to Processing is performed to obtain the optimal right endpoint of the characteristic fuzzy model of the infectious disease patient population ,in, express around The optimal right endpoint of the constructed s-th feature fuzzy model on the j-th feature;

[0031] Step 3.3: Construct the re-fuzzy fitness value of the feature fuzzy model and use the genetic algorithm to and After processing, the characteristic fuzzy model of the final H infectious disease patient population is obtained ,and , represents the left endpoint of the feature fuzzy model with aggressive estimation, represents the conservatively estimated left endpoint of the feature fuzzy model, represents the right endpoint of the feature fuzzy model with conservative estimation, Represents the right endpoint of the feature fuzzy model with aggressive estimation.

[0032] The characteristic fuzzy model construction method for infectious disease patients described in the present invention is also characterized in that: in step 3.1, the left endpoint fitness value is constructed using formula (10):

[0033] (10)

[0034] In formula (10), express Around The constructed s-th feature fuzzy model is the left endpoint of the j-th feature and serves as an individual in the genetic population; represents a function for calculating the number of patients with an infectious disease, Indicates that it is located The jth feature of the nth infectious patient in represents the number of features that satisfy the inequality; A function representing the accuracy of the calculated feature fuzzy model;

[0035] Furthermore, step 3.2 is to construct the right endpoint fitness value using formula (11):

[0036] (11)

[0037] In formula (11), express Around The constructed s-th feature fuzzy model is the right endpoint of the j-th feature and serves as an individual in the genetic population.

[0038] Furthermore, the step 3.3 is to construct the re-fuzzy fitness value using formula (12):

[0039] (12)

[0040] In formula (12), express Around The parameter set of the constructed s-th feature fuzzy model on the j-th feature is taken as an individual in the genetic population. The parameter set corresponding to each individual includes: Around The constructed s-th feature fuzzy model is based on the left endpoint of the j-th feature with radical estimation ; Around The constructed s-th feature fuzzy model is based on the j-th feature with a conservative estimate of the left endpoint ; Around The right endpoint of the conservative estimate of the constructed s-th feature fuzzy model on the j-th feature ; Around The right endpoint of the radical estimate of the constructed s-th feature fuzzy model on the j-th feature ; represents the optimal individual, and st represents the constraint relationship.

[0041] The electronic device of the present invention includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the feature fuzzy model construction method, and the processor is configured to execute the program stored in the memory.

[0042] The present invention provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program, which is characterized in that when the computer program is run by a processor, the steps of the feature fuzzy model construction method are executed.

[0043] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0044] 1. The present invention uses type II fuzzy sets as the model framework of the patient population, which has stronger interpretability than the clusters generated by the clustering algorithm, and can convey more patient information than type I fuzzy granular data. The fuzzy description of the characteristics of the patient group finally obtained makes the infected population model modeled using this method easier to understand.

[0045] 2. The present invention introduces the idea of ​​density estimation data mining. By comprehensively considering the representativeness, uniqueness and noise evaluation of patients, the selected patient representatives are more representative, more diverse and less affected by noise than the data prototypes generated by the clustering algorithm.

[0046] 3. The present invention fully considers the amount of information contained in the infected population feature model and the accuracy of the description, and introduces a genetic algorithm to optimize these two indicators at the same time. Facing the patients themselves, the final model has high coverage and strong specificity. A good balance is achieved between the interpretability and accuracy of the patient population. And by splitting the problem, the construction process of the feature fuzzy model is split into three sub-steps, which effectively reduces the construction time of the infected population feature model. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] Figure 1 A flowchart constructed for the medical patient population model of the present invention;

[0048] Figure 2 A schematic diagram of the fuzzy type II granular data modeling process used in the present invention;

[0049] Figure 3 It is a schematic diagram of the fuzzy type II granularity data of the present invention at the jth feature. DETAILED DESCRIPTION

[0050] In this embodiment, a method for modeling infected populations based on a type II fuzzy granular model is used to model populations based on medical patient information. First, medical-related information of patients is collected to construct a comprehensive set of patient features. Then, the representativeness, uniqueness and noise resistance scores of each patient are calculated to select patient representatives with high total scores. Finally, a fuzzy interval type II granular model is constructed around the patient representatives. The model contains fuzzy feature descriptions of multiple patients. Figure 1 , the modeling method is carried out in the following steps:

[0051] Step 1: Collect the medical records, examination reports, medical history, living habits and other characteristic information of infectious disease patients, and build an infectious disease patient dataset , Indicates Infectious disease patients The jth feature in is the total number of patients with infectious diseases, is the total number of characteristics of each infectious disease patient;

[0052] Step 2: Select representatives of infectious disease patients; Figure 2 As shown in the first step, representative patients are selected from the patient-feature dataset.

[0053] Step 2.1: Use DBSCAN algorithm to Divide into Infectious disease clusters , and count the number of infectious disease patients in each infectious disease patient cluster and the number of patients with noise-related infectious diseases ,in, Indicates Clusters of infectious patients, express The number of patients with infectious diseases ,lie in Any qth infectious disease patient in , ;

[0054] Step 2.2: Calculate the number of infectious disease patient representatives ,in, is the selected ratio. In this embodiment, the selected ratio p is 0.25. The selected ratio can be adjusted according to specific needs. Thus, according to formula (3), the first Infectious disease clusters The number of infectious disease patient representatives in ;like Figure 2 In , the number of patient representatives H is 3, which also means that three patient models will be constructed in the end.

[0055] (3)

[0056] Step 2.3, calculation After sorting the Euclidean distance between any two infectious disease patients in descending order, select the first The mean of the Euclidean distances is The density radius , thus obtaining The density radius ;

[0057] Step 2.4: Calculate using formula (5) The qth infectious disease patient Representative score of ; Representativeness is represented by the number of neighbor data of the patient. The larger the value, the larger the neighbor and the better the representativeness.

[0058] (5)

[0059] In formula (5), Represents a positive activation function; express The qth infectious patient With the hth infectious disease patient The Euclidean distance between

[0060] Step 2.5: Calculate according to formula (6) The qth infectious disease patient Uniqueness score After normalization, the normalized patient uniqueness score is obtained ; Uniqueness is measured by the distance from other highly representative data, the greater the distance, the higher the uniqueness;

[0061] (6)

[0062] In formula (6), express The hth infectious disease patient Representative scores of

[0063] Step 2.6: Calculate according to formula (8) The qth infectious disease patient The non-noise score :

[0064] (8)

[0065] In formula (8), express Middle Infectious disease patients No. Features, express Middle The minimum value of the features, express Middle The maximum value of features;

[0066] Step 2.7: Calculate according to formula (9) Middle Infectious disease patients Final candidate score :

[0067] (9)

[0068] Step 3: Based on the infectious disease patient representatives selected in step 2, a fuzzy model of the infectious disease patient population is constructed; Figure 2 As shown in the second step, the patient representation is used to construct a fuzzy model of the patient.

[0069] Step 3.1: Construct the left endpoint fitness value of the feature fuzzy model and use the left endpoint genetic algorithm to Processing is performed to obtain the optimal left endpoint of the characteristic fuzzy model of infectious disease patients ,in, express Around The optimal left endpoint of the constructed s-th feature fuzzy model on the j-th feature; the optimization goal is composed of two indicators: the large number of patients covered by the model and the accurate range of features described by the model;

[0070] (10)

[0071] In formula (10), express Around The constructed s-th feature fuzzy model takes the left endpoint on the j-th feature as an individual in the population; A calculation function representing the number of infectious disease patients, Indicates that it is located in The jth feature of the nth infectious disease patient in the infectious disease patient cluster, represents the number of features that satisfy the inequality; The calculation function that represents the accuracy of the feature fuzzy model; Figure 3 As shown, the genetic algorithm will be on the jth feature Expand to the left .

[0072] Step 3.2: Construct the right endpoint fitness value of the feature fuzzy model and use the right endpoint genetic algorithm to Processing is performed to obtain the optimal right endpoint of the characteristic fuzzy model of infectious disease patients ,in, express around The optimal right endpoint of the constructed s-th feature fuzzy model on the j-th feature;

[0073] (11)

[0074] In formula (11), express Around The constructed s-th feature fuzzy model takes the right endpoint of the j-th feature as an individual in the population; Figure 3 As shown, the algorithm will be on the jth feature Expand to the right , and Together they constitute the initial interval data. The method of expanding the left and right intervals separately makes full use of the independence of interval data and effectively improves the convergence speed of the genetic algorithm.

[0075] Step 3.3: Construct the refuzzy fitness value of the feature fuzzy model and use the refuzzification genetic algorithm to and After processing, the characteristic fuzzy model of the final H infectious disease patients is obtained ,and , represents the left endpoint of the model with aggressive estimation, represents the conservatively estimated left endpoint of the model, represents the right end point of the model with conservative estimation, Represents the right endpoint of the model with radical estimation. The individual fitness formula of the genetic algorithm is as follows (12).

[0076] (12)

[0077] In formula (12), express Around The parameter set of the constructed s-th feature fuzzy model on the j-th feature is taken as an individual in the population. The parameter set corresponding to each individual includes: Around The constructed s-th feature fuzzy model is based on the left endpoint of the j-th feature with radical estimation ; Around The constructed s-th feature fuzzy model is based on the j-th feature with a conservative estimate of the left endpoint ; Around The right endpoint of the conservative estimate of the constructed s-th feature fuzzy model on the j-th feature ; Around The right endpoint of the radical estimate of the constructed s-th feature fuzzy model on the j-th feature ; represents the optimal individual, and st represents the constraint relationship. Figure 3 As shown, the algorithm will be on the jth feature Expand to and ,Will Expand to and ,The re-fuzzification of the two interval endpoints promotes the type 1 fuzzy data to type 2 fuzzy data, which contains more patient characteristic information.

[0078] As a generative model, the fuzzy model provides a fuzzy characteristic description of the infectious disease population. Unlike the discriminant model which has only a single use, the model can provide doctors with multiple assistances in diagnosing and understanding the characteristics of the disease.

[0079] For example, we now have a fuzzy model for a certain infected population, where feature j represents the temperature feature of the population. If we want to judge the question feature of this population at a 60% confidence level, the model will output: Under conservative estimates, when the temperature is When the patient's temperature is within this range, the probability that the patient is ill is above 60%. When a radical estimate is required, the probability that the patient is ill is greater than 60%. .

[0080] For example, the model can output a judgment on the possibility of a suspected patient being sick. Now there is a new suspected patient with a body temperature of T. First, the infected population model to which the patient belongs is found by the proximity principle. , we can judge by its body temperature characteristics, when The probability of disease is In this range, when The probability of disease is Within this range.

[0081] In this embodiment, an electronic device includes a memory and a processor, wherein the memory is used to store a program that supports the processor to execute the above method, and the processor is configured to execute the program stored in the memory.

[0082] In this embodiment, a computer-readable storage medium stores a computer program on the computer-readable storage medium, and the computer program executes the steps of the above method when executed by a processor.

[0083] In summary, the method of the present invention not only classifies patients into several groups, but also provides a fuzzy feature description of the patient group, which will significantly improve the amount of information and interpretability contained in the model, which is beneficial for doctors to understand the characteristics of infectious diseases and the characteristics of susceptible populations, and provide auxiliary information for the diagnosis of future suspected patients. At the same time, this method also fully considers the statistical errors and noise problems inherent in big data, can effectively eliminate the impact of these data on the accuracy of the final model, and further improve the robustness of the fuzzy model of infected population characteristics.

Claims

1. A method for constructing a characteristic fuzzy model for patients with infectious diseases, characterized in that: The steps include: Step 1: Collect various characteristics of infectious disease patients and build an infectious disease patient dataset , Indicates Infectious disease patients The jth feature in is the total number of patients with infectious diseases, is the total number of characteristics of each infectious disease patient; Step 2: Select representatives of infectious disease patients; Step 2.1: Use DBSCAN algorithm to Divide into Infectious disease clusters , and count the number of infectious disease patients in each infectious disease patient cluster and the number of patients with noise-related infectious diseases ,in, Indicates Clusters of infectious patients, express The number of patients with infectious diseases ,lie in Any qth infectious disease patient in , ; Step 2.2: Calculate the number of infectious disease patient representatives ,in, is the selected ratio, so according to formula (3) Infectious disease clusters The number of infectious disease patient representatives in ; (3) Step 2.3, calculation After sorting the Euclidean distance between any two infectious disease patients in descending order, select the first The mean of the Euclidean distances is The density radius , thus obtaining The density radius ; Step 2.4: Calculate using formula (5) The qth infectious disease patient Representative score of ; (5) In formula (5), Represents a positive activation function; express The qth infectious patient With the hth infectious disease patient The Euclidean distance between Step 2.5: Calculate according to formula (6) The qth infectious disease patient Uniqueness score After normalization, the normalized patient uniqueness score is obtained ; (6) In formula (6), express The hth infectious disease patient Representative scores of Step 2.6: Calculate according to formula (8) The qth infectious disease patient The non-noise score : (8) In formula (8), express Middle Infectious disease patients No. Features, express Middle The minimum value of the features, express Middle The maximum value of features; Step 2.7: Calculate according to formula (9) Middle Infectious disease patients Final candidate score : (9) Step 2.8: The final candidate scores of all infectious disease patients in the Infectious disease patients Representatives of infectious disease patients , thus obtaining Infectious disease clusters Representatives of infectious disease patients ,in, express The sth infectious disease patient representative; Step 3: Based on the infectious disease patient representatives selected in step 2, a characteristic fuzzy model of the infectious disease patient population is constructed; Step 3.1: Construct the left endpoint fitness value of the feature fuzzy model and use the genetic algorithm to Processing is performed to obtain the optimal left endpoint of the characteristic fuzzy model of the infectious disease patient population ,in, express Around The optimal left endpoint of the constructed s-th feature fuzzy model on the j-th feature; Step 3.2: Construct the right endpoint fitness value of the feature fuzzy model and use the genetic algorithm to Processing is performed to obtain the optimal right endpoint of the characteristic fuzzy model of the infectious disease patient population ,in, express around The optimal right endpoint of the constructed s-th feature fuzzy model on the j-th feature; Step 3.3: Construct the re-fuzzy fitness value of the feature fuzzy model and use the genetic algorithm to and After processing, the characteristic fuzzy model of the final H infectious disease patient population is obtained ,and , represents the left endpoint of the feature fuzzy model with aggressive estimation, represents the conservatively estimated left endpoint of the feature fuzzy model, represents the right endpoint of the feature fuzzy model with conservative estimation, Represents the right endpoint of the feature fuzzy model with aggressive estimation.

2. A method for constructing a characteristic fuzzy model for infectious disease patients according to claim 1, characterized in that: In step 3.1, the left endpoint fitness value is constructed using formula (10): (10) In formula (10), express Around The constructed sth feature fuzzy model is the left endpoint of the jth feature and serves as an individual in the genetic population; represents a function for calculating the number of patients with an infectious disease, Indicates that it is located The jth feature of the nth infectious patient in represents the number of features that satisfy the inequality; Represents the function for calculating the accuracy of the feature fuzzy model.

3. A method for constructing a characteristic fuzzy model for infectious disease patients according to claim 2, characterized in that: The step 3.2 is to construct the right endpoint fitness value using formula (11): (11) In formula (11), express Around The constructed s-th feature fuzzy model is the right endpoint of the j-th feature and serves as an individual in the genetic population.

4. The method for constructing a characteristic fuzzy model for infectious disease patients according to claim 2, characterized in that: The step 3.3 is to construct the re-fuzzy fitness value using formula (12): (12) In formula (12), express Around The parameter set of the constructed s-th feature fuzzy model on the j-th feature is taken as an individual in the genetic population. The parameter set corresponding to each individual includes: Around The constructed s-th feature fuzzy model is based on the left endpoint of the j-th feature with radical estimation ; Around The constructed s-th feature fuzzy model is based on the j-th feature with a conservative estimate of the left endpoint ; Around The right endpoint of the conservative estimate of the constructed s-th feature fuzzy model on the j-th feature ; Around The right endpoint of the radical estimate of the constructed s-th feature fuzzy model on the j-th feature ; represents the optimal individual, and st represents the constraint relationship.

5. An electronic device, comprising a memory and a processor, characterized in that: The memory is used to store a program that supports the processor to execute the feature fuzzy model construction method described in any one of claims 1 to 4, and the processor is configured to execute the program stored in the memory.

6. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the feature fuzzy model construction method described in any one of claims 1 to 4 are executed.

Images