Intelligent screening system based on abnormal changes in human sweat

By applying cosine similarity algorithm and correlation network model in the intelligent screening system, the problem of slow sweat analysis speed in the prior art is solved, rapid disease screening and auxiliary diagnosis are achieved, and analysis efficiency and accuracy are improved.

CN111710409BActive Publication Date: 2024-11-29吾征智能技术(北京)有限公司
View PDF 10 Cites 0 Cited by

Patent Information

Application Number
CN202010476532.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2020-05-29
Publication Date
2024-11-29
Estimated Expiration
2040-05-29

AI Technical Summary

Technical Problem

The prior art cannot improve sweat analysis speed by establishing evaluation screening models and cosine similarity algorithms, resulting in complex analysis processes, long waiting times and high resource consumption.

Method used

An intelligent screening system based on abnormal changes in human sweat is proposed, including a collection module, a network model establishment module, a screening standard establishment module and a screening module. The semantic similarity calculation is carried out through the cosine similarity algorithm to establish the abnormal change feature information entry A network model with disease characteristic information to achieve rapid screening and auxiliary diagnosis.

Benefits of technology

By using the cosine similarity algorithm to calculate semantic similarity, risk assessment of abnormal sweat changes and disease screening are realized, possible diseases are speculated, and auxiliary diagnosis services are served, and sweat analysis is accelerated, and the accuracy of auxiliary diagnosis of diseases is improved.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN111710409B_ABST
    Figure CN111710409B_ABST
Patent Text Reader

Abstract

The present invention proposes an intelligent screening system based on abnormal changes in human sweat. It includes: a collection module, which is used to collect abnormal change feature information corresponding to abnormal changes in human sweat; a network model establishment module, which is used to establish an association network model between abnormal change feature information entries and corresponding disease feature information; a screening standard establishment module, which is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entries and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data; a screening module, which is used to screen the abnormal change feature information of the sweat to be screened according to the feature information screening standard. The present invention uses the cosine similarity algorithm to perform semantic similarity calculation, realizes risk assessment of abnormal sweat changes and disease screening, infers possible diseases, and serves as auxiliary diagnosis.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to an intelligent screening system based on abnormal changes in human sweat. Background Art

[0002] Under normal circumstances, the sweat that people excrete is conditional. For example, the sweat excreted by healthy people is colorless, but when the sweat excreted by people has an abnormal color, it is a signal of disease. At present, the international and domestic methods of observing sweat to identify diseases are mainly achieved through sensor devices. This sensor device is generally composed of a flexible sensor and a microprocessor. It can be attached to the skin to stimulate the sweat glands to secrete sweat, and then measure the molecules and ions in the sweat based on the relevant electrical signals.

[0003] However, the traditional method of sweat analysis based on sensors collecting sweat components has a complex analysis process, long waiting time, and high resource consumption. Therefore, there is an urgent need for a new sweat analysis method that can quickly and conveniently assist in the diagnosis of diseases while consuming a small amount of resources.

[0004] The above contents are only used to assist in understanding the technical solution of the present invention and do not constitute an admission that the above contents are prior art. Summary of the invention

[0005] In view of this, the present invention proposes an intelligent screening system based on abnormal changes in human sweat, aiming to solve the technical problem that the prior art cannot improve the sweat analysis speed by establishing an evaluation screening model and a cosine similarity algorithm.

[0006] The technical solution of the present invention is achieved in this way:

[0007] In one aspect, the present invention provides an intelligent screening system based on abnormal changes in human sweat, the intelligent screening system based on abnormal changes in human sweat comprising:

[0008] A collection module, used for collecting abnormal change characteristic information corresponding to abnormal changes in human sweat;

[0009] A network model building module is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information;

[0010] A screening standard establishment module is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data;

[0011] The screening module is used to obtain the characteristic information of abnormal changes in sweat to be screened, and screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information.

[0012] On the basis of the above technical scheme, preferably, the acquisition module includes a natural language processing module, which is used to process the abnormal change characteristic information corresponding to the abnormal changes in human sweat through natural language, and obtain attribute information of the abnormal change characteristic information of sweat, wherein the attribute information includes: the location, color, quantity and smell of the abnormal changes in sweat.

[0013] On the basis of the above technical scheme, preferably, the network model building module includes a clustering module, which is used to cluster the attribute information of the abnormal change characteristic information of sweat, and obtain the clustered attribute information as the abnormal change characteristic information entry, wherein the abnormal change characteristic information entry includes: occurrence time, occurrence quantity, sweat state, sweat color and sweat odor, and establish a corresponding relationship between each abnormal change characteristic information entry.

[0014] Based on the above technical solutions, preferably, the network model building module also includes an associated network model module, which is used to obtain corresponding common disease characteristic information based on abnormal change characteristic information entries, and establish an associated network model based on the abnormal change characteristic information entries and the corresponding common disease characteristic information.

[0015] On the basis of the above technical scheme, preferably, the screening standard establishment module includes a similarity calculation module, which is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, and verify and modify the association network model according to the similarity to obtain a new association network model.

[0016] On the basis of the above technical solution, preferably, the screening standard establishment module includes a cosine similarity algorithm unit, and the cosine similarity algorithm is:

[0017] ;

[0018] in, Represents cosine similarity, vector a and vector b Represents abnormal change characteristic information entries and corresponding disease characteristic information, c Representation vector a and vector b The distance between , x 1 and y 1 is a vector a The coordinates ofx 2 and y 2 is a vector b The coordinates of .

[0019] On the basis of the above technical solution, preferably, the screening module includes a report generating module, which is used to obtain the characteristic information of abnormal changes in sweat to be screened, screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information, and generate a corresponding report according to the screening results.

[0020] More preferably, the intelligent screening device based on abnormal changes in human sweat includes:

[0021] A collection unit, used for collecting abnormal change characteristic information corresponding to abnormal changes in human sweat;

[0022] A network model building unit is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information;

[0023] A screening standard establishing unit is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data;

[0024] The screening unit is used to obtain the characteristic information of abnormal changes in sweat to be screened, and screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information.

[0025] The intelligent screening system based on abnormal changes in human sweat of the present invention has the following beneficial effects compared with the prior art:

[0026] By using the cosine similarity algorithm to calculate semantic similarity, the risk assessment of abnormal sweat changes and disease screening can be achieved, possible diseases can be inferred, and auxiliary diagnosis services can be provided;

[0027] By using the cosine similarity algorithm to calculate semantic similarity, the speed of sweat analysis can be accelerated and the accuracy of disease auxiliary diagnosis can be improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0029] Figure 1 It is a structural block diagram of the first embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention;

[0030] Figure 2 It is a structural block diagram of a second embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention;

[0031] Figure 3 It is a structural block diagram of a third embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention;

[0032] Figure 4 This is a structural block diagram of a fourth embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention;

[0033] Figure 5 It is a structural block diagram of a fifth embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention;

[0034] Figure 6 This is a structural block diagram of the intelligent screening system based on abnormal changes in human sweat of the present invention, and an intelligent screening device based on abnormal changes in human sweat. DETAILED DESCRIPTION

[0035] The following will be combined with 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.

[0036] like Figure 1 As shown, Figure 1 The structure block diagram of the first embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention includes: a collection module 10, a network model building module 20, a screening standard building module 30 and a screening module 40.

[0037] A collection module 10 is used to collect abnormal change characteristic information corresponding to abnormal changes in human sweat;

[0038] The network model building module 20 is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information;

[0039] A screening standard establishing module 30 is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data;

[0040] The screening module 40 is used to obtain the characteristic information of abnormal changes in sweat to be screened, and screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information.

[0041] Furthermore, if Figure 2 As shown, based on the above embodiments, a second embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention is proposed. In this embodiment, the acquisition module 10 also includes:

[0042] The natural language processing module 101 is used to process the abnormal change characteristic information corresponding to the abnormal change of human sweat through natural language to obtain attribute information of the abnormal change characteristic information of sweat, wherein the attribute information includes: the location, color, quantity and smell of the abnormal change of sweat;

[0043] It should be understood that, first, the system will intensively sample and obtain multiple abnormal change feature information presented by abnormal changes in human sweat, and then, through natural language processing, express the feature information of each abnormal change in sweat accordingly according to its characteristics to obtain the attribute information of the abnormal change feature information of sweat, for example: the location, color, quantity and smell of abnormal changes in sweat.

[0044] Furthermore, if Figure 3 As shown, based on the above embodiments, a third embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention is proposed. In this embodiment, the network model building module 20 also includes:

[0045] The clustering module 201 is used to cluster the attribute information of the abnormal change characteristic information of sweat, obtain the clustered attribute information as the abnormal change characteristic information entry, and the abnormal change characteristic information entry includes: appearance time, appearance quantity, sweat state, sweat color and sweat odor, and establish a corresponding relationship between each abnormal change characteristic information entry.

[0046] The association network model module 202 is used to obtain corresponding common disease characteristic information according to the abnormal change characteristic information entry, and to establish an association network model according to the abnormal change characteristic information entry and the corresponding common disease characteristic information.

[0047] It should be understood that the system clusters the description information of these characteristic points of sweat, gathers significant categories, and establishes associations between them. For example: the location of abnormal sweat changes (for example, sweating on the forehead, sweating on the chest, etc.), the time of occurrence (for example, sweating at night, sweating in the morning, and cold sweat, etc.), the number of occurrences (for example, sweating profusely or no sweat, etc.), and the state of sweat (for example, sweating continuously or sweating continuously, etc.), the color of sweat (for example, yellow sweat, red sweat, etc.), the smell of sweat (for example, the smell of sweat is sour, the sweat has a feather smell, the sweat has a fishy smell, etc.), etc.

[0048] It should be understood that the specific clustering operation steps here include: constructing a sweat feature point fusion degree matrix according to the linear fusion similarity, and obtaining the preset value of the sweat feature point according to the sweat feature point fusion degree matrix K clusters; according to the preset values ​​of sweat feature points K The clusters are judged for clustering convergence conditions. If any of the conditions is met, new user normalized sweat feature points can be obtained. K The cluster convergence conditions are judged for each cluster. If any of the conditions is not met, the step of constructing the sweat feature point fusion matrix by linear fusion similarity is performed. The cluster convergence conditions include: no abnormal change feature points are reallocated to different classes, or the new cluster center is the same as the original cluster center.

[0049] It should be understood that, later, the system will cluster all the description information of all the abnormal sweat change feature points to obtain multiple segmentations or entries to form a dictionary of abnormal sweat change feature information. M The descriptive information includes "profuse sweating (excessive sweating), fever, thirst, drinking a lot of water, red tongue, yellow tongue coating", etc.; another feature point information N The description information includes "excessive sweating, pale complexion, constant dripping, cold limbs, confusion, and extremely weak pulse", etc., and another feature point information The description information includes "excessive sweating, dizziness, general weakness, obvious hunger", etc. Then, all the description information can be clustered through the above steps to obtain the following "excessive sweating" entries of sweat: "excessive sweating (or heavy sweating), fever, thirst, drinking a lot of water, red tongue, yellow tongue coating, pale complexion, constant sweating, cold limbs, confusion, very weak pulse, dizziness, general weakness, obvious hunger", etc., and classified into the category of "abnormal changes in sweat, quantity", thus forming a dictionary of abnormal changes in sweat feature information.

[0050] It should be understood that common disease characteristic information (symptoms or signs) that corresponds to abnormal sweat change characteristic information will be collected later. For example, "hot sweat on the palms" is classified as "abnormal sweat change, location", so the corresponding disease may be "Yin deficiency". The main symptoms (or signs) of Yin deficiency generally include "hot sweat on the palms and soles, hot five hearts, or hot sweat on the palms and soles, afternoon hot flashes, night sweats, red cheeks", etc., so a segmentation or entry library of common diseases and corresponding symptoms (or signs) can be established.

[0051] It should be understood that after the steps of the above embodiment, the system will construct a network model for associating abnormal sweat change feature information with diseases and disease symptoms (or signs) (for example, the relationship between abnormal sweat color feature information and diseases and disease symptom feature information, etc.). Here, the main method for identifying and marking segmentation or phrases in the text description of abnormal sweat changes is: matching the segmentation or phrase in the segmentation or phrase with the segmentation or phrase in the disease word library, if the match is successful, marking the segmentation or phrase as a disease segmentation or phrase, otherwise, marking the segmentation or phrase as a description segmentation or phrase; here, the description segmentation or phrase refers to the disease symptom description segmentation or phrase; the disease word library contains the segmentation or phrases of each disease symptom (or sign) in the nine systems of the human motor system, digestive system, respiratory system, urinary system, reproductive system, endocrine system, immune system, nervous system, and circulatory system.

[0052] Furthermore, if Figure 4 As shown, based on the above embodiments, a fourth embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention is proposed. In this embodiment, the screening standard establishment module 30 also includes:

[0053] The similarity calculation module 301 is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, and verify and modify the association network model according to the similarity to obtain a new association network model.

[0054] Cosine similarity algorithm unit 302, the cosine similarity algorithm is:

[0055] ;

[0056] in, Represents cosine similarity, vector a and vector b Represents abnormal change characteristic information entries and corresponding disease characteristic information, c Representation vector a and vector b The distance between , x 1 and y 1 is a vector a The coordinates of x 2 and y 2 is a vector b The coordinates of .

[0057] It should be understood that the system will use the cosine similarity algorithm to calculate the similarity between the semantic relationship between the abnormal sweat change feature information segmentation or entry and the disease symptom (or sign) segmentation or entry, and perform internal verification of the association network model.

[0058] It should be immediately noted that in the rectangular coordinate system, the vector a Use coordinates Indicates that the vector b Use coordinates Indicates that the vector a and vector b The length in rectangular coordinates is , ,vector a and vector b We use the vector c ,so , from which we can get the formula of the cosine function in 2D space:

[0059] ;

[0060] It should be understood that, assuming there are 3 items, item1, item2 and item3, the vector representations are:

[0061] Item1 , item2 , item3 , that is, 3 points in the five-dimensional space. The distance between item1 and item2, and the distance between item2 and item3 are calculated using the Euclidean distance formula, which are:

[0062] item1-item2=7, item2-item3=1;

[0063] The cosine value of the angle between item1 and item2 is calculated using the cosine function:

[0064] ;

[0065] Use the cosine function to calculate the cosine value of the angle between item2 and item3:

[0066] ;

[0067] It can be concluded that item1 and item2 have a small similarity, and the distance between them is large (the distance is 7), while item2 and item3 have a large similarity, and the distance between them is small (the distance is 1).

[0068] The following example illustrates how to use cosine similarity to calculate the similarity between two terms:

[0069] List all words; encode words; vectorize word frequency; apply cosine function to measure the similarity between two words;

[0070] Patient's main complaint feature item A: yellow sweat, liver area discomfort, jaundice, frequent edema, fatigue, difficulty in urination and defecation, etc.;

[0071] Disease cirrhosis symptoms entry B: yellow sweat, discomfort in the liver area, jaundice, ascites, edema, bleeding, fatigue, yellow urine, etc.;

[0072] List all the words, put listA and listB in a set, and get: set = {yellow sweat, liver discomfort, jaundice, frequent edema, fatigue, urination and defecation problems, ascites, bleeding, yellow urine}. Convert the above set to dict, with the key being the word in the set and the value being the position of the word in the set, i.e., 'yellow urine':1;

[0073] Encode listA and listB, converting each word into the position where it appears in the set;

[0074] Perform oneHot encoding on listAcode and listBcode, and calculate the number of occurrences of each word;

[0075] After obtaining the frequency vectors of the two terms, it becomes a matter of calculating the cosine value of the angle between the two vectors. The larger the value, the higher the similarity. According to the cosine similarity, terms A and B are very similar.

[0076] It should be understood that the basic standard for screening the characteristic information of sweat and the characteristic information of disease symptoms (or signs) is obtained according to the above matching degree; at the same time, the linear fusion similarity is clustered and collaborative filtering recommendation is performed. For example, there is only one participle or phrase related to sweat in terms A and B: "yellow sweat". Through the clustering operation of linear fusion similarity, it belongs to the category of sweat color and is classified into this category.

[0077] Furthermore, if Figure 5 As shown, based on the above embodiments, a fifth embodiment of the intelligent screening system based on abnormal changes in human sweat of the present invention is proposed. In this embodiment, the screening module 40 also includes:

[0078] The report generation module 401 is used to obtain the characteristic information of abnormal changes in sweat to be screened, screen the characteristic information of abnormal changes in sweat to be screened according to the characteristic information screening criteria, and generate a corresponding report according to the screening results.

[0079] It should be understood that this screening standard is used for screening and analysis of abnormal sweat conditions within the model application domain. When establishing the sweat characteristic information screening model, this embodiment, in addition to considering the sweat dictionary model, also takes into account the spatial position relationship between various property points of sweat (including sweat color, sweat odor, sweat state, sweat amount, sweating time, sweating parts of the body, etc.), so as to more accurately construct a classification model for sweat characteristic classification, thereby more accurately classifying sweat properties.

[0080] It should be understood that the following is an embodiment, for example: a disease, whose symptoms (or signs) include primary symptoms (or signs) and secondary symptoms (or signs). The primary symptom (or sign) refers to the symptom (or sign) that the disease will definitely manifest; the secondary symptom (or sign) refers to the complications that may be caused by the disease. Therefore, for the calculation of the semantic similarity of the terms, the primary and secondary status of the symptoms (or signs) need to be considered, and they cannot be generalized. For example, the main symptom (or sign) of a cold is fever, while sweating, coughing, etc. are secondary symptoms (or signs). For the descriptions of the conditions of two cold patients, "fever, sweating, and coughing" and "fever, no sweating, no coughing", when calculating the semantic similarity of the abnormal sweat change feature information, if the primary and secondary are not considered, the descriptions of the two conditions, then their similarity is very low, but in fact, these two conditions may both be colds, which are very similar.

[0081] It should be noted that the above is only an example and does not constitute any limitation to the technical solution of the present application.

[0082] It is not difficult to find out from the above description that this embodiment proposes an intelligent screening system based on abnormal changes in human sweat, including: a collection module, which is used to collect abnormal change feature information corresponding to abnormal changes in human sweat; a network model establishment module, which is used to establish an association network model between abnormal change feature information entries and corresponding disease feature information; a screening standard establishment module, which is used to establish a cosine similarity algorithm, calculate the similarity between abnormal change feature information entries and corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish feature information screening standards based on the similarity data; a screening module, which is used to screen the abnormal change feature information of the sweat to be screened according to the feature information screening standards. This embodiment uses the cosine similarity algorithm to perform semantic similarity calculations to achieve risk assessment of abnormal sweat changes and disease screening, infer possible diseases, and provide auxiliary diagnosis services.

[0083] In addition, the embodiment of the present invention also provides an intelligent screening device based on abnormal changes in human sweat. Figure 6 As shown, the intelligent screening device based on abnormal changes in human sweat includes: a collection unit 10, a network model building unit 20, a screening standard building unit 30 and a screening unit 40.

[0084] A collection unit 10 is used to collect abnormal change characteristic information corresponding to abnormal changes in human sweat;

[0085] The network model building unit 20 is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information;

[0086] The screening standard establishing unit 30 is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change characteristic information entry and the corresponding disease characteristic information by the cosine similarity algorithm, obtain similarity data, and establish a characteristic information screening standard according to the similarity data;

[0087] The screening unit 40 is used to obtain the characteristic information of abnormal changes in sweat to be screened, and screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information.

[0088] In addition, it should be noted that the device embodiments described above are merely illustrative and do not limit the scope of protection of the present invention. In practical applications, technicians in this field can select some or all of the modules according to actual needs to achieve the purpose of the present embodiment, and no restrictions are made here.

[0089] In addition, for technical details not fully described in this embodiment, reference can be made to the intelligent screening system based on abnormal changes in human sweat provided in any embodiment of the present invention, and will not be repeated here.

[0090] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An intelligent screening system based on abnormal changes in human sweat, characterized in that: The intelligent screening system based on abnormal changes in human sweat includes: A collection module, used for collecting abnormal change characteristic information corresponding to abnormal changes in human sweat; The acquisition module includes a natural language processing module, which is used to process the abnormal change characteristic information corresponding to the abnormal change of human sweat through natural language, and obtain the attribute information of the abnormal change characteristic information of sweat, wherein the attribute information includes: the location, color, quantity and smell of the abnormal change of sweat; A network model building module is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information; The network model building module includes a clustering module, which is used to cluster the attribute information of the abnormal change characteristic information of sweat, obtain the clustered attribute information as the abnormal change characteristic information entry, and the abnormal change characteristic information entry includes: appearance time, appearance quantity, sweat state, sweat color and sweat odor, and establish a corresponding relationship between each abnormal change characteristic information entry; The network model building module also includes an associated network model module, which is used to obtain corresponding common disease characteristic information according to the abnormal change characteristic information entry, and to build an associated network model according to the abnormal change characteristic information entry and the corresponding common disease characteristic information; Clustering the description information of the characteristic points of sweat, clustering significant categories, and establishing associations between the description information, wherein the description information includes the location, time, amount, state, color, and smell of the sweat where the abnormal change of sweat occurs; The clustering operation steps include: constructing a fusion degree matrix of sweat feature points according to the linear fusion similarity, obtaining a preset value of sweat feature point clusters according to the sweat feature point fusion degree matrix; judging clustering convergence conditions according to the preset value of sweat feature point clusters, and obtaining new user normalized sweat feature points when any one of the clustering convergence conditions is met; when all the clustering convergence conditions are not met, executing the step of constructing a fusion degree matrix of sweat feature points according to the linear fusion similarity, and the clustering convergence conditions include: no abnormal change feature points are reallocated to different classes, or the new cluster center is the same as the original cluster center; A screening standard establishment module is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data; The screening standard establishment module includes a similarity calculation module, which is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, and verify and modify the association network model according to the similarity to obtain a new association network model; The screening criteria establishment module includes a cosine similarity algorithm unit, and the cosine similarity algorithm is: ; in, Represents cosine similarity, vector a and vector b Represents abnormal change characteristic information entries and corresponding disease characteristic information, c Representation vector a and vector b The distance between , x 1 and y 1 is a vector a The coordinates of x 2 and y 2 is a vector b The coordinates of The cosine similarity algorithm is used to calculate the similarity between the semantic relationship between the abnormal sweat change feature information segmentation or entry and the disease symptom segmentation or entry, and the association network model is internally verified; By getting the formula for the cosine function in 2D space: ; Suppose there are three items, item1, item2 and item3, represented by vectors: Item1 , item2 , item3 , that is, three points in the five-dimensional space, use the Euclidean distance formula to calculate the distance between item1 and itme2. The larger the distance, the smaller the similarity. A screening module is used to obtain the characteristic information of abnormal changes in sweat to be screened, and screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information; The screening module includes a report generating module, which is used to obtain the characteristic information of abnormal changes in sweat to be screened, screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information, and generate a corresponding report according to the screening results.

2. An intelligent screening device based on abnormal changes in human sweat, characterized in that: The intelligent screening device based on abnormal changes in human sweat includes: A collection unit, used for collecting abnormal change characteristic information corresponding to abnormal changes in human sweat; The collection unit includes a natural language processing module, which is used to process the abnormal change characteristic information corresponding to the abnormal change of human sweat through natural language, and obtain attribute information of the abnormal change characteristic information of sweat, wherein the attribute information includes: the location, color, quantity and smell of the abnormal change of sweat; A network model building unit is used to cluster the abnormal change characteristic information, extract abnormal change characteristic information entries from the clustered abnormal change characteristic information, obtain corresponding disease characteristic information according to the abnormal change characteristic information entries, and establish an association network model between the abnormal change characteristic information entries and the corresponding disease characteristic information; The network model building unit includes a clustering module, which is used to cluster the attribute information of the abnormal change characteristic information of sweat, obtain the clustered attribute information as the abnormal change characteristic information entry, and the abnormal change characteristic information entry includes: appearance time, appearance quantity, sweat state, sweat color and sweat odor, and establish a corresponding relationship between each abnormal change characteristic information entry; The network model building unit also includes an associated network model module, which is used to obtain corresponding common disease characteristic information according to the abnormal change characteristic information entry, and to build an associated network model according to the abnormal change characteristic information entry and the corresponding common disease characteristic information; Clustering the description information of the characteristic points of sweat, clustering significant categories, and establishing associations between the description information, wherein the description information includes the location, time, amount, state, color, and smell of the sweat where the abnormal change of sweat occurs; The clustering operation steps include: constructing a fusion degree matrix of sweat feature points according to the linear fusion similarity, obtaining a preset value of sweat feature point clusters according to the sweat feature point fusion degree matrix; judging clustering convergence conditions according to the preset value of sweat feature point clusters, and obtaining new user normalized sweat feature points when any one of the clustering convergence conditions is met; when all the clustering convergence conditions are not met, executing the step of constructing a fusion degree matrix of sweat feature points according to the linear fusion similarity, and the clustering convergence conditions include: no abnormal change feature points are reallocated to different classes, or the new cluster center is the same as the original cluster center; A screening standard establishing unit is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, obtain similarity data, and establish a feature information screening standard based on the similarity data; The screening standard establishment unit includes a similarity calculation module, which is used to establish a cosine similarity algorithm, calculate the similarity between the abnormal change feature information entry and the corresponding disease feature information through the cosine similarity algorithm, and verify and modify the association network model according to the similarity to obtain a new association network model; The screening standard establishing unit includes a cosine similarity algorithm unit, and the cosine similarity algorithm is: ; in, Represents cosine similarity, vector a and vector b Represents abnormal change characteristic information entries and corresponding disease characteristic information, c Representation vector a and vector b The distance between , x 1 and y 1 is a vector a The coordinates of x 2 and y 2 is a vector b The coordinates of The cosine similarity algorithm is used to calculate the similarity between the semantic relationship between the abnormal sweat change feature information segmentation or entry and the disease symptom segmentation or entry, and the association network model is internally verified; By getting the formula for the cosine function in 2D space: ; Suppose there are three items, item1, item2 and item3, represented by vectors: Item1 , item2 , item3 , that is, three points in the five-dimensional space, use the Euclidean distance formula to calculate the distance between item1 and itme2. The larger the distance, the smaller the similarity. A screening unit, used for obtaining characteristic information of abnormal changes in sweat to be screened, and screening the characteristic information of abnormal changes in sweat to be screened according to a screening standard of the characteristic information; The screening module includes a report generating module, which is used to obtain the characteristic information of abnormal changes in sweat to be screened, screen the characteristic information of abnormal changes in sweat to be screened according to the screening criteria of the characteristic information, and generate a corresponding report according to the screening results.

Citation Information

Patent Citations

  • Abnormal behavior detection method and system based on big-data association rule mining

    CN105843947A

  • Recommendation method based on label semantic normalization

    CN106599037A

  • Modeling method for parallel smart case recommendation model

    CN107656952A

  • Method and system for carrying out automatic disease code conversion on basis of semantic approximate matching algorithm

    CN108446260A

  • Data query system based on quality text similarity

    CN109241246A