CN113545771B - Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure - Google Patents

Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure Download PDF

Info

Publication number
CN113545771B
CN113545771B CN202110782848.4A CN202110782848A CN113545771B CN 113545771 B CN113545771 B CN 113545771B CN 202110782848 A CN202110782848 A CN 202110782848A CN 113545771 B CN113545771 B CN 113545771B
Authority
CN
China
Prior art keywords
training
neighbor
signal
plantar pressure
module
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202110782848.4A
Other languages
Chinese (zh)
Other versions
CN113545771A (en
Inventor
曹军义
赵欢
王瑞雪
殷志成
谢俊枭
雷亚国
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Xian Jiaotong University
Original Assignee
Xian Jiaotong University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Xian Jiaotong University filed Critical Xian Jiaotong University
Priority to CN202110782848.4A priority Critical patent/CN113545771B/en
Publication of CN113545771A publication Critical patent/CN113545771A/en
Application granted granted Critical
Publication of CN113545771B publication Critical patent/CN113545771B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/103—Detecting, measuring or recording devices for testing the shape, pattern, colour, size or movement of the body or parts thereof, for diagnostic purposes
    • A61B5/1036—Measuring load distribution, e.g. podologic studies
    • A61B5/1038—Measuring plantar pressure during gait
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/40—Detecting, measuring or recording for evaluating the nervous system
    • A61B5/4076—Diagnosing or monitoring particular conditions of the nervous system
    • A61B5/4082—Diagnosing or monitoring movement diseases, e.g. Parkinson, Huntington or Tourette
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/68—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient
    • A61B5/6801—Arrangements of detecting, measuring or recording means, e.g. sensors, in relation to patient specially adapted to be attached to or worn on the body surface
    • A61B5/6802—Sensor mounted on worn items
    • A—HUMAN NECESSITIES
    • A61—MEDICAL OR VETERINARY SCIENCE; HYGIENE
    • A61B—DIAGNOSIS; SURGERY; IDENTIFICATION
    • A61B5/00—Measuring for diagnostic purposes; Identification of persons
    • A61B5/72—Signal processing specially adapted for physiological signals or for diagnostic purposes
    • A61B5/7235—Details of waveform analysis
    • A61B5/7264—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems
    • A61B5/7267—Classification of physiological signals or data, e.g. using neural networks, statistical classifiers, expert systems or fuzzy systems involving training the classification device
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2413—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on distances to training or reference patterns
    • G06F18/24147—Distances to closest patterns, e.g. nearest neighbour classification
    • G—PHYSICS
    • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
    • G16H—HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA
    • G16H50/00—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics
    • G16H50/20—ICT specially adapted for medical diagnosis, medical simulation or medical data mining; ICT specially adapted for detecting, monitoring or modelling epidemics or pandemics for computer-aided diagnosis, e.g. based on medical expert systems
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/08—Feature extraction
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/12—Classification; Matching

Landscapes

  • Health & Medical Sciences (AREA)
  • Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Public Health (AREA)
  • Biomedical Technology (AREA)
  • Physics & Mathematics (AREA)
  • Medical Informatics (AREA)
  • General Health & Medical Sciences (AREA)
  • Pathology (AREA)
  • Veterinary Medicine (AREA)
  • Animal Behavior & Ethology (AREA)
  • Surgery (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • Biophysics (AREA)
  • Heart & Thoracic Surgery (AREA)
  • Neurology (AREA)
  • Data Mining & Analysis (AREA)
  • Neurosurgery (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Physiology (AREA)
  • Theoretical Computer Science (AREA)
  • Evolutionary Computation (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Fuzzy Systems (AREA)
  • Databases & Information Systems (AREA)
  • Epidemiology (AREA)
  • Primary Health Care (AREA)
  • Developmental Disabilities (AREA)
  • General Engineering & Computer Science (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Biology (AREA)
  • Mathematical Physics (AREA)
  • Psychiatry (AREA)
  • Signal Processing (AREA)
  • Dentistry (AREA)
  • Oral & Maxillofacial Surgery (AREA)
  • Medical Treatment And Welfare Office Work (AREA)
  • Measurement Of The Respiration, Hearing Ability, Form, And Blood Characteristics Of Living Organisms (AREA)

Abstract

An integrated K-nearest neighbor quantitative Parkinson disease diagnosis system based on plantar pressure comprises a signal acquisition preprocessing module, a feature extraction module, a training learning output module and a control module, wherein the signal acquisition preprocessing module is used for acquiring a signal; the signal acquisition and preprocessing module acquires pressure signals of a Parkinson disease person and a healthy person, and preprocessing the pressure signals to obtain signal fragments; the feature extraction module is used for extracting the plantar pressure features, firstly standardizing, secondly determining the intercepting step length, and then determining the plantar pressure feature type and feature set; the training learning output module trains the K neighbor classification model, acquires partial features from the training feature set as input of a K neighbor classification algorithm to form a plurality of K neighbor base classifiers, calculates the output performance of each K neighbor base classifier, calculates the weight of each K neighbor base classifier, completes integrated learning, and completes quantitative diagnosis of the Parkinson's disease by taking the label with the maximum output probability as the health state corresponding to the plantar pressure signal; the invention can efficiently and intelligently complete the quantitative diagnosis of the disease degree of the Parkinson disease patient.

Description

Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure
Technical Field
The invention belongs to the technical field of Parkinson disease diagnosis, and particularly relates to an integrated K-nearest neighbor Parkinson disease quantitative diagnosis system based on plantar pressure.
Background
With the aggravation of the aging phenomenon of the society, the number of common senile patients such as the Parkinson disease and the like is increased greatly. As a third killer for middle-aged and elderly people after tumor and cardiovascular and cerebrovascular diseases, the diagnosis and treatment of the Parkinson disease are widely concerned by experts in various fields. However, early symptoms are hidden, and when typical clinical symptoms do not appear yet, the patient has nervous system degenerative changes, and the disease course is progressive, so that quantitative diagnosis is needed to provide support for clinical treatment. The international unified Parkinson disease rating scale which is commonly used is a UPDRS scale, normal, mild, moderate, severe and extreme severe degrees of the Parkinson disease are evaluated by observing patients by doctors, and the methods have certain credibility and effectiveness on the evaluation of the severe degree of the Parkinson disease, but have strong subjectivity and can not accurately and uniformly evaluate the Parkinson disease. The plantar pressure in the normal walking process is used as observation data, has the advantages of non-invasion, small interference, high comfort and the like, and can become basic characteristic source data for evaluating the severity and the development of the Parkinson disease. As a classic mode classification method, the K nearest neighbor method can identify and classify data according to existing empirical data. The method has the advantages of relatively simple structure and algorithm, and the like, and is widely applied to multiple fields of voice, images and the like.
However, the conventional K-nearest neighbor algorithm has two problems when applied to quantitative diagnosis of parkinson's disease. (1) The human healthy individuals have obvious self-difference and are influenced by living habits, occupation, education degrees and the like, different individuals show different changes of sole pressure, and in addition, pathological representation is also complicated and changeable under the influence of environmental factors, so that the difference of healthy groups and Parkinson disease patient groups is submerged in the self-difference of the individuals, and the diagnosis precision is reduced; (2) In the actual clinical diagnosis process, the number of samples of healthy patients and patients with mild diseases is huge, however, the gait data sample volume of patients with serious diseases is extremely rare, and the data with extremely unbalanced distribution cannot learn to obtain a reliable K neighbor classification model which can sufficiently represent the whole fault data distribution, so that the low diagnosis precision is finally caused, and the good diagnosis effect is difficult to obtain.
Disclosure of Invention
In order to overcome the defects of the prior art, the invention aims to provide an integrated K-nearest neighbor quantitative diagnosis system for Parkinson's disease based on plantar pressure, which can efficiently and intelligently complete quantitative diagnosis of the disease degree of Parkinson's disease patients in a wearable background.
In order to achieve the purpose, the invention adopts the technical scheme that:
an integrated K-nearest neighbor quantitative diagnosis system for Parkinson disease based on plantar pressure comprises a signal acquisition preprocessing module, a feature extraction module and a training learning output module;
the signal acquisition preprocessing module: pressure signals of different positions of soles of a Parkinson patient and a healthy person in the normal walking process are obtained to form a sole pressure signal set
Figure BDA0003157667060000021
Preprocessing the plantar pressure signal to obtain a signal segment;
a feature extraction module: carrying out gait feature extraction, firstly standardizing the signal segment, and secondly determining the step length I of the intercepted signal segment train Then determining the type and the characteristic set of the plantar pressure characteristic;
training and learning output module: training the K neighbor classification model, and training the feature set for multiple times
Figure BDA0003157667060000031
Partial acquired features are used as input of a K neighbor classification algorithm to form a plurality of different K neighbor base classifiers; calculating the output performance g of each K neighbor base classifier m And F1 m ；
G according to each base classifier m And F1 m Calculating the weight of each K neighbor base classifier, completing ensemble learning, and after training is completed, taking the label with the maximum output probability as a plantar pressure signal x n And according to the health state, completing the quantitative diagnosis of the Parkinson disease.
The pressure signals of different positions of the sole in the signal acquisition preprocessing module comprise: heel, inner and outer arch, inner and outer sole and sole pressure change information at the position of thumb 8.
The signal processing process in the feature extraction module specifically comprises the following steps:
firstly, the plantar pressure signals of the signal segments are standardized, and the signals of all pressure sensors on the soles are divided by the body weight of a tested person to construct a standard signal set
Figure BDA0003157667060000032
From plantar pressure signals of each individual separately
Figure BDA0003157667060000033
In step I train Intercepting N in Sample fragments of length, constituting a set of sample fragments
Figure BDA0003157667060000034
Wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0003157667060000035
is the jth sample fragment of the nth individual and contains N in A data point;
respectively obtaining the pressure mean value characteristics of the sample segments
Figure BDA0003157667060000036
Wherein the content of the first and second substances,
Figure BDA0003157667060000037
is the ith point in the jth sample segment of the nth individual and the coefficient of variation characteristics of the sample segment
Figure BDA0003157667060000038
Then extracting the asymmetric characteristics representing the coordination of the left and right lower limbs,
Figure BDA0003157667060000039
forming training feature sets
Figure BDA00031576670600000310
The training and learning process in the training and learning output module is specifically as follows:
training feature set
Figure BDA00031576670600000311
Randomly assigning features in the set to different training sets of base classifiers
Figure BDA0003157667060000041
Forming m different K neighbor base classifiers as the input characteristics of the mth K neighbor classification algorithm, and calculating the output performance of each K neighbor base classifier, g m And F 1 Is to evaluate the effective index for the unbalanced data classification algorithm,
Figure BDA0003157667060000042
for multi-classification, the real class label is 1 as the concerned class, TP is that the predicted class is consistent with the real class, TN is also accurate when the predicted class is the class, FN is that other classes are predicted as the concerned class, FP is that other classes are predicted as the classes;
calculating the weight value of each K neighbor base classifier in the ensemble learning process
Figure BDA0003157667060000043
And (4) taking the weight of each base classifier as an integration standard in the integrated learning process to finish integrated K nearest neighbor classification and realize quantitative diagnosis of the Parkinson disease.
The beneficial effects of the invention are as follows:
the invention integrates a K nearest neighbor classification algorithm, directly extracts disease degree characteristics from plantar pressure signals, overcomes individual difference of characteristics of the traditional method, and extracts characteristics with pathological differentiation; and secondly, the diagnosis precision is finally improved for the quantitative diagnosis of the Parkinson disease by utilizing the integrated K nearest neighbor learning method.
Drawings
Fig. 1 is a signal processing flow chart of the present invention.
FIG. 2 is a graph comparing diagnostic results of examples of the present invention.
Detailed Description
The invention is described in further detail below with reference to the figures and examples.
Referring to fig. 1, the integrated K-nearest neighbor quantitative Parkinson disease diagnosis system based on plantar pressure comprises a signal acquisition preprocessing module, a feature extraction module and a training output module;
the signal acquisition preprocessing module: acquiring pressure signals of different positions of soles of a Parkinson patient and healthy old people in the normal walking process, comprising the following steps of: the change information of the plantar pressure at the heel, the inner heel, the outer heel, the inner arch, the outer sole and the thumb 8 forms a plantar pressure signal set
Figure BDA0003157667060000051
Preprocessing the plantar pressure signal, and removing a head effect and a tail effect by using signal data in a walking stabilization stage to obtain a signal segment;
a feature extraction module: performing gait feature extraction, firstly normalizing the signal segments to eliminate partial individual differences, and secondly determining the step length I of intercepting the signal segments train And then determining the plantar pressure characteristic type and the characteristic set, specifically:
firstly, the plantar pressure signals of the signal segments are standardized, and the signals of all pressure sensors on the soles are divided by the body weight of a tested person to construct a standard signal set
Figure BDA0003157667060000052
From plantar pressure signals of each individual separately
Figure BDA0003157667060000053
In step I train Intercepting N in Sample fragments of length, constituting a set of sample fragments
Figure BDA0003157667060000054
Wherein, the first and the second end of the pipe are connected with each other,
Figure BDA0003157667060000055
is the jth sample fragment of the nth individual and contains N in A data point;
respectively obtaining the pressure mean value characteristics of the sample segments
Figure BDA0003157667060000056
Wherein the content of the first and second substances,
Figure BDA0003157667060000057
is the ith point in the jth sample segment of the nth individual and the coefficient of variation characteristics of the sample segment
Figure BDA0003157667060000058
Then extracting the asymmetric characteristics representing the coordination of the left and right lower limbs,
Figure BDA0003157667060000059
forming training feature sets
Figure BDA00031576670600000510
A training output module: training the K neighbor classification model, and performing multiple times of training feature set
Figure BDA00031576670600000511
Partial acquired features are used as input of a K neighbor classification algorithm to form a plurality of different K neighbor base classifiers; the output performance of each K-neighbor base classifier is calculated,
Figure BDA0003157667060000061
g according to each base classifier m And F1 m Calculating to obtain the weight of each K neighbor base classifier, completing ensemble learning, and after training is completed, taking the label with the maximum output probability as a plantar pressure signal x n And (3) completing quantitative diagnosis of the Parkinson disease according to the corresponding health state, specifically comprising the following steps:
training feature set
Figure BDA0003157667060000062
Randomly assigning features in the set to different training sets of base classifiers
Figure BDA0003157667060000063
Forming m different K neighbor base classifiers as the input characteristics of the mth K neighbor classification algorithm, and calculating the output performance of each K neighbor base classifier due to g m And F 1 Is to evaluate the effective index for the unbalanced data classification algorithm,
Figure BDA0003157667060000064
for multi-classification, for example, in classification of four classes, 0,1,2,3, the true class label is 1 as the concerned class, TP is that the predicted class is consistent with the true class, that is, the prediction class label is 1, tn is accurate when the predicted class is the class label of the predicted class, that is, the prediction class 0 is 0, and the prediction class 2 is 2, etc.; FN is to predict other classes as the class of interest, i.e. possibly class labeled 0,2,3 is to predict class 1, FP is to predict other classes as class, i.e. possibly class labeled 0 is to predict class labeled 2 or 3;
calculating the weight value of each K neighbor base classifier in the integrated learning process
Figure BDA0003157667060000065
And (3) taking the weight of each base classifier as an integration standard in the integrated learning process to finish the integrated K nearest neighbor classification and realize the quantitative diagnosis of the Parkinson disease.
Example (b): the invention is further described by intelligent quantitative diagnosis of disease degree in parkinson's patients of different disease degree.
The gait data set of Parkinson disease comprises gait data of 71 healthy old people and 91 Parkinson disease people, wherein 55 of 91 Parkinson disease people are H&Y2 stage, 27 names of Pah&Y2.5, 9 being H&Y3 phase, containing about 4000 data points per sample; about 80% of all samples were used for model training, and the remaining samples were used for model accuracy testing. Make it possible toThe invention is used for diagnosing the gait data set of the Parkinson disease, and the step length I of the sample segment is intercepted aiming at the data set train At 20, using the present invention, a training accuracy of 96.91% and a testing accuracy of 95.02% can be achieved (a single K-nearest neighbor method yields a training accuracy of 96.71% and a testing accuracy of 94.15%). To verify the effectiveness of the present invention, the diagnostic results of the present invention are compared to those based on time phase characteristics and integrated SVM methods. In a conventional gait phase feature-based diagnostic method (method 1), the extracted features include: gait cycle, swing phase time, stance phase time, double support phase time, and the coefficient of variation and left and right foot asymmetry of these features; then, the characteristics are respectively used as input, and the intelligent quantitative diagnosis of the Parkinson disease with different disease degrees is completed by utilizing the integrated K nearest neighbor method provided by the invention; in the diagnostic method (method 2) of the integrated SVM, the plantar pressure characteristic provided by the invention is taken as input, and the disease degree quantitative diagnosis of the Parkinson disease is completed by using the integrated SVM method.
Compared with the three diagnosis results shown in fig. 2, it can be seen from the comparison result of fig. 2 that the training precision of the time phase feature is 82.65% and the testing precision is 76.03%, which are far lower than the training precision and the testing precision of the present invention, which shows that the pressure feature of the present invention contains more disease state information than the traditional time phase feature, has better distinguishing capability, overcomes the problem that the traditional time phase feature is difficult to be classified by an intelligent algorithm, and obtains higher diagnosis precision. The training precision of the integrated SVM method is 73.24%, and the testing precision is 72.59%, which are lower than the training precision and the testing precision of the method, so that the method can overcome the imbalance of empirical data and realize accurate diagnosis of diseases in different degrees.
Through the analysis of the specific treatment process of the quantitative diagnosis of the Parkinson disease degree and the diagnosis result based on the plantar pressure signal characteristics, the plantar pressure characteristics of the invention have better disease degree identification capability, but the invention can overcome the problem that the distribution of samples of patients with different disease degrees is unbalanced and is not beneficial to the classification of an intelligent algorithm, and the high-efficiency and reliable quantitative intelligent diagnosis is carried out on the Parkinson disease patients with different disease degrees by combining an integrated K neighbor classification method based on the provided plantar pressure signal characteristics, so that higher diagnosis precision is obtained.

Claims (1)

1. An integrated K-nearest neighbor quantitative diagnosis system for Parkinson disease based on plantar pressure is characterized in that: the system comprises a signal acquisition preprocessing module, a feature extraction module and a training learning output module;
the signal acquisition preprocessing module: pressure signals of different positions of soles of a Parkinson patient and a healthy person in the normal walking process are obtained to form a sole pressure signal set
Figure FDA0003659085040000011
Preprocessing the plantar pressure signal to obtain a signal segment;
a feature extraction module: carrying out gait feature extraction, firstly standardizing the signal segment, and secondly determining the step length I of intercepting the signal segment train Then determining the type and the characteristic set of the plantar pressure characteristic;
the training and learning output module: training the K neighbor classification model, and performing multiple times of training feature set
Figure FDA0003659085040000012
Partial acquired features are used as input of a K neighbor classification algorithm to form a plurality of different K neighbor base classifiers; calculating the output performance g of each K neighbor base classifier m And F1 m (ii) a G according to each base classifier m And F1 m Calculating to obtain the weight of each K neighbor base classifier, completing ensemble learning, and after training is completed, using the label with the maximum output probability as the plantar pressure signal x n Completing quantitative diagnosis of the Parkinson disease according to the corresponding health state;
the pressure signals of different positions of the sole in the signal acquisition preprocessing module comprise: heel, inner and outer heels, inner and outer foot arches, inner and outer soles and sole pressure change information at the position of the thumb 8;
the signal processing process in the feature extraction module is specifically as follows:
firstly, the plantar pressure signals of the signal segments are standardized, and the signals of all pressure sensors on the soles are divided by the body weight of a tested person to construct a standard signal set
Figure FDA0003659085040000021
From plantar pressure signals of each individual separately
Figure FDA0003659085040000022
In step I train Intercepting N in Sample fragments of length, constituting a set of sample fragments
Figure FDA0003659085040000023
Wherein the content of the first and second substances,
Figure FDA0003659085040000024
is the jth sample fragment of the nth individual and contains N in A data point;
respectively obtaining the pressure mean value characteristics of the sample segments
Figure FDA0003659085040000025
Wherein, the first and the second end of the pipe are connected with each other,
Figure FDA0003659085040000026
is the ith point in the jth sample segment of the nth individual and the coefficient of variation characteristics of the sample segment
Figure FDA0003659085040000027
Then extracting the asymmetric characteristics representing the coordination of the left and right lower limbs,
Figure FDA0003659085040000028
forming training feature sets
Figure FDA0003659085040000029
The training and learning process in the training and learning output module is specifically as follows:
training feature set
Figure FDA00036590850400000210
Randomly assigning features in the graph to different training sets of base classifiers
Figure FDA00036590850400000211
Forming m different K neighbor base classifiers as the input characteristics of the mth K neighbor classification algorithm, and calculating the output performance of each K neighbor base classifier, g m And F 1 Is to evaluate the effective index for the unbalanced data classification algorithm,
Figure FDA00036590850400000212
for multi-classification, the real class label is 1 as the concerned class, TP is the predicted class which is consistent with the real class, TN is the predicted class which is also accurate, FN is the class which predicts other classes as the concerned class, FP is the class which predicts other classes as classes;
calculating the weight value of each K neighbor base classifier in the integrated learning process
Figure FDA00036590850400000213
And (3) taking the weight of each base classifier as an integration standard in the integrated learning process to finish the integrated K nearest neighbor classification and realize the quantitative diagnosis of the Parkinson disease.
CN202110782848.4A 2021-07-12 2021-07-12 Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure Active CN113545771B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110782848.4A CN113545771B (en) 2021-07-12 2021-07-12 Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110782848.4A CN113545771B (en) 2021-07-12 2021-07-12 Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure

Publications (2)

Publication Number Publication Date
CN113545771A CN113545771A (en) 2021-10-26
CN113545771B true CN113545771B (en) 2022-10-28

Family

ID=78103015

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110782848.4A Active CN113545771B (en) 2021-07-12 2021-07-12 Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure

Country Status (1)

Country Link
CN (1) CN113545771B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN116473514B (en) * 2023-03-29 2024-02-23 西安电子科技大学广州研究院 Parkinson disease detection method based on plantar pressure self-adaptive directed space-time graph neural network

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109717833A (en) * 2018-11-26 2019-05-07 中国科学院软件研究所 A kind of neurological disease assistant diagnosis system based on human motion posture
WO2020244771A1 (en) * 2019-06-06 2020-12-10 NEC Laboratories Europe GmbH Diversity-aware weighted majority vote classifier for imbalanced datasets
WO2021063935A1 (en) * 2019-09-30 2021-04-08 F. Hoffmann-La Roche Ag Prediction of disease status

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104834888A (en) * 2014-12-04 2015-08-12 龙岩学院 Abnormal gait identification method capable of facilitating screening Parkinsonism
US20190365287A1 (en) * 2018-05-30 2019-12-05 Industry-Academic Cooperation Foundation, Dankook University Apparatus and method for gait type classification using pressure sensor of smart insole
CN113057623B (en) * 2021-03-30 2022-09-06 上海理工大学 Gait analysis system for screening flat foot

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109717833A (en) * 2018-11-26 2019-05-07 中国科学院软件研究所 A kind of neurological disease assistant diagnosis system based on human motion posture
WO2020244771A1 (en) * 2019-06-06 2020-12-10 NEC Laboratories Europe GmbH Diversity-aware weighted majority vote classifier for imbalanced datasets
WO2021063935A1 (en) * 2019-09-30 2021-04-08 F. Hoffmann-La Roche Ag Prediction of disease status

Also Published As

Publication number Publication date
CN113545771A (en) 2021-10-26

Similar Documents

Publication Publication Date Title
Vásquez-Correa et al. Multimodal assessment of Parkinson's disease: a deep learning approach
CN108492877B (en) Cardiovascular disease auxiliary prediction method based on DS evidence theory
CN105930663A (en) Parkinson's disease early diagnosis method
CN113796852B (en) Diabetes foot prediction method based on gradient lifting decision tree model algorithm
CN114999629A (en) AD early prediction method, system and device based on multi-feature fusion
CN109934179B (en) Human body action recognition method based on automatic feature selection and integrated learning algorithm
CN115563484A (en) Street greening quality detection method based on physiological awakening identification
Singh et al. Short unsegmented PCG classification based on ensemble classifier
CN111798980B (en) Complex medical biological signal processing method and device based on deep learning network
CN113545771B (en) Integrated K-neighbor quantitative Parkinson disease diagnosis system based on plantar pressure
CN114358194A (en) Gesture tracking based detection method for abnormal limb behaviors of autism spectrum disorder
CN115527608A (en) Intestinal age prediction method and system
Zhang et al. Patient-specific seizure prediction from scalp EEG using vision transformer
CN111932522A (en) Alzheimer's disease classifier based on brain imaging big data deep learning
Zhao et al. Accurate identification of Parkinson’s disease by distinctive features and ensemble decision trees
CN112444500B (en) Alzheimer's disease intelligent detection device based on spectrum
CN114299996A (en) AdaBoost algorithm-based speech analysis method and system for key characteristic parameters of symptoms of frozen gait of Parkinson's disease
CN107045624A (en) Electroencephalogram signal preprocessing and classifying method based on maximum weighted cluster
CN113796873A (en) Wearable dynamic electrocardiosignal classification method and system
CN116473514B (en) Parkinson disease detection method based on plantar pressure self-adaptive directed space-time graph neural network
CN113317780A (en) Abnormal gait detection method based on long-time and short-time memory neural network
CN116509335A (en) Gradual evolution automatic sleep stage-dividing method
CN116484290A (en) Depression recognition model construction method based on Stacking integration
Haque et al. Evaluating performance of machine learning models for diabetic sensorimotor polyneuropathy severity classification using biomechanical signals during gait
CN112384128A (en) Method for measuring pain intensity

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant