CN117648572B - Cognitive assessment method and system based on virtual reality and ensemble learning - Google Patents
Cognitive assessment method and system based on virtual reality and ensemble learning Download PDFInfo
- Publication number
- CN117648572B CN117648572B CN202410122491.0A CN202410122491A CN117648572B CN 117648572 B CN117648572 B CN 117648572B CN 202410122491 A CN202410122491 A CN 202410122491A CN 117648572 B CN117648572 B CN 117648572B
- Authority
- CN
- China
- Prior art keywords
- representing
- samples
- sample
- cognitive
- classifier
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 84
- 230000001149 cognitive effect Effects 0.000 title claims abstract description 48
- 230000008569 process Effects 0.000 claims abstract description 27
- 238000012549 training Methods 0.000 claims abstract description 7
- 238000007781 pre-processing Methods 0.000 claims abstract description 5
- 238000000513 principal component analysis Methods 0.000 claims abstract description 4
- 208000010877 cognitive disease Diseases 0.000 claims description 15
- 238000012360 testing method Methods 0.000 claims description 14
- 230000003044 adaptive effect Effects 0.000 claims description 10
- 230000006870 function Effects 0.000 claims description 9
- 230000002194 synthesizing effect Effects 0.000 claims description 9
- 230000006399 behavior Effects 0.000 claims description 7
- 230000003920 cognitive function Effects 0.000 claims description 7
- 238000011156 evaluation Methods 0.000 claims description 7
- 238000005070 sampling Methods 0.000 claims description 6
- 238000013145 classification model Methods 0.000 claims description 4
- 238000005457 optimization Methods 0.000 claims description 4
- 208000028698 Cognitive impairment Diseases 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 230000008859 change Effects 0.000 claims description 3
- 208000027061 mild cognitive impairment Diseases 0.000 claims description 3
- 230000035945 sensitivity Effects 0.000 claims description 3
- 238000012795 verification Methods 0.000 claims description 3
- 230000019771 cognition Effects 0.000 claims 1
- 238000013473 artificial intelligence Methods 0.000 abstract description 2
- 230000006999 cognitive decline Effects 0.000 description 5
- 238000010586 diagram Methods 0.000 description 5
- 238000001514 detection method Methods 0.000 description 3
- 238000010801 machine learning Methods 0.000 description 3
- 238000005259 measurement Methods 0.000 description 3
- 230000009286 beneficial effect Effects 0.000 description 2
- 238000002591 computed tomography Methods 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 238000002595 magnetic resonance imaging Methods 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 239000000090 biomarker Substances 0.000 description 1
- 210000004556 brain Anatomy 0.000 description 1
- 230000003930 cognitive ability Effects 0.000 description 1
- 230000036992 cognitive tasks Effects 0.000 description 1
- 230000001010 compromised effect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000012252 genetic analysis Methods 0.000 description 1
- 230000001771 impaired effect Effects 0.000 description 1
- 230000004630 mental health Effects 0.000 description 1
- 230000006996 mental state Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000007479 molecular analysis Methods 0.000 description 1
- 238000002610 neuroimaging Methods 0.000 description 1
- 238000010855 neuropsychological testing Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000036632 reaction speed Effects 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 230000000638 stimulation Effects 0.000 description 1
Landscapes
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The invention relates to the field of artificial intelligence, in particular to a cognitive assessment method and a system based on virtual reality and ensemble learning. The cognitive assessment method based on virtual reality and ensemble learning carries out preprocessing on data; carrying out balance treatment on a minority class sample by using a KM-BorderlineMOTE algorithm; performing dimension reduction on the features by using a principal component analysis algorithm, and taking the features with the accumulated variance contribution rate of more than 95% as classification features; selecting N-class base classifiers for model training; and integrating the output results of the N-class base classifier by using a self-adaptive comprehensive weight updating integrated learning algorithm. The cognitive assessment system based on virtual reality and integrated learning comprises a user subsystem, a virtual supermarket subsystem and a cognitive assessment subsystem. Compared with the prior art, the method can better capture the characteristics and rules of the data and improve the generalization capability and classification performance of the model; the system improves the ecological effectiveness of the cognitive assessment process.
Description
Technical Field
The invention relates to the field of artificial intelligence, in particular to a cognitive assessment method and a system based on virtual reality and ensemble learning.
Background
The cognitive decline of the elderly has a wide and profound impact on both life and society. First, cognitive decline in the elderly may lead to difficulties in activities of daily living such as hypomnesis, inattention, and impaired decision making; secondly, social isolation is an important problem brought by cognitive decline, such as forgetting the names of friends and relatives, failing to participate in deep dialogue or failing to follow social rules, leading them to gradually disjoint from society, thereby affecting mental health and quality of life; in addition, cognitive decline may also result in their financial management being compromised, such as failure to properly manage bank accounts, pay bills, etc.
Cognitive assessment is a process of screening for cognitive function in an individual, aimed at detecting cognitive dysfunction or risk of cognitive dysfunction, as well as providing early intervention and treatment. However, current cognitive assessment has certain limitations:
(1) Neuroimaging techniques such as MRI (magnetic resonance imaging) and CT (computed tomography) can be used for cognitive assessment, but multiple detections expose the patient to radiation and provide structural information that does not allow direct measurement of functional changes or identify early cognitive decline;
(2) The spatial resolution of electroencephalogram EEG is relatively low, detailed images of the internal structures of the brain cannot be provided, and interference from external noise may lead to inaccurate results;
(3) Biomarker detection typically involves complex molecular or genetic analysis, requiring expertise and equipment;
(4) Neuropsychological tests are widely used in the detection of cognitive disorders at present, including montreal cognitive assessment (MOCA), mini Mental State Examination (MMSE), etc., and these classical conventional psychological tests are often used in a quiet test environment when used for cognitive assessment, but our usual cognitive activities often occur in a more complex environment, so that the measurement result of this method is relatively coarse, the ecological effectiveness is poor, and the actual problems encountered by patients in real environments are completely incomparable.
Virtual reality is an analog simulation technology, which allows participants to utilize handles, helmets, microphones, and the like to perform immersive interactions with objects in a virtual environment, and can provide vivid, real-world analog environments for the participants, thereby better simulating daily cognitive tasks. The rich information and data contained in the virtual environment need to be calculated and analyzed by a machine learning method so as to discover information related to cognitive assessment. Such information includes data of various aspects of user behavior, perceptual feedback, interaction patterns, etc., which constitute a complex information network. The patterns and their associations are extracted from these massive data by sophisticated machine learning algorithms. The method can be used for improving the design and performance optimization of the virtual environment, and is also beneficial to customizing the user experience and improving the accuracy and efficiency of cognitive assessment. Therefore, in order to solve the above-mentioned problems, there is a need for a cognitive assessment method and system that is integrated with a real complex environment for assessing the cognitive ability of the elderly.
Disclosure of Invention
The invention aims to solve the limitations of the existing cognitive assessment method and the existing cognitive assessment system, and utilizes the thinking of virtual reality and machine learning to provide the cognitive assessment method and the system based on virtual reality and integrated learning in the environment of multi-sensory stimulation and feedback under the condition of guaranteeing exercise safety.
In order to achieve the above object, the present invention provides a cognitive assessment method based on virtual reality and ensemble learning, comprising the steps of:
s100, dividing a final data set into a training set, a test set and a verification set according to the proportion of 8:1:1 after preprocessing the data;
s200, aiming at the problem of unbalanced quantity of two types of samples, providing a KM-Borderlin eSMOTE algorithm to up-sample a few types of samples;
s300, reducing the dimension of the features by using a principal component analysis algorithm, sorting the features in descending order according to the variance contribution rate after reducing the dimension, and taking the features with the accumulated variance contribution rate more than 95% as classification features;
s400, selecting an N-type base classifier for model training;
s500, integrating output results of the N-class base classifier by utilizing a self-adaptive comprehensive weight updating integrated learning algorithm to obtain a final classification class.
The KM-BorderlineSMOTE algorithm in the S200 comprises the following steps:
s210, calculating K neighbors of all minority class samples by using K neighbor algorithms, and dividing minority class sample properties at the same time: k neighbors of the ith minority class sample containMinority class samples and->Multiple types of samples, if->The properties of the minority class samples are classified as Safe; if->The properties of the minority class samples are divided into Danger; if->The properties of the minority class samples are classified as Noise;
s220, synthesizing a new sample:
s221, synthesizing new samples by using a Borderline-SMOTE algorithm on minority samples with Danger properties, wherein the number of required manual generation is as follows:
in the method, in the process of the invention,minority class samples representing properties DangerNumber of (A)>The number of the minority class samples with the property of Safe is represented, and M represents the total number of the samples in the data set;
s222, for minority samples with Safe properties, firstly, clustering the minority samples with Safe properties by using an improved K-Medoids algorithm, and dividing a data set into K clusters; then setting sampling multiplying powerThe method comprises the following steps:
in the method, in the process of the invention,sample rate for the i-th cluster, +.>Represents the cluster center of the ith cluster, +.>Represents the j-th sample in the i-th cluster,/>Representation sample->To cluster center->K represents the number of clusters, +.>Representing the number of small samples in the ith cluster;
s223, synthesizing a new sample according to the sampling multiplying power:
in the method, in the process of the invention,representing a new sample of the composition, +.>And->Representing a minority class of samples and randomly selected samples in its K-nearest neighbor;
s230, combining the synthesized new sample with the original sample data set to obtain a final data set.
Wherein, the modified K-Medoids algorithm in S222 includes: the clustering criterion function is introduced to the K-Medoids algorithm, so that the influence of abnormal values is reduced, meanwhile, the dividing result is closer to the cluster center point, and the improved optimization objective function is as follows:
in the method, in the process of the invention,expressed as a change to the clustering criterion function, M represents the total number of samples in the dataset, <>Represents standard deviation of kth cluster, +.>Represents the cluster center of the kth cluster, +.>Represents the j-th sample in the k-th cluster,/>Representation sample->To cluster center->Is a distance of (3).
The self-adaptive comprehensive weight updating integrated learning algorithm in S500 includes: self-adaptive comprehensive weight updating formula based on relative accuracy, measuring performance index AUC of classification model and unbalanced classification index G-mean：
In the method, in the process of the invention,representing an adaptive comprehensive weight updating formula; />And->Respectively representing the maximum error rate and the minimum error rate in all the base classifiers; />Representing the error rate of the kth base classifier; />The area not covered by the ROC curve representing the kth basis classifier, i.e. +.>Wherein->Representing the area covered under the ROC curve of the kth basis classifier; />Representing the maximum of the uncovered areas of all the basis classifier ROC curves; />Representing the minimum of uncovered areas of all base classifier ROC curves; />G-mean value representing kth basis classifier, i.e.>Where TPR represents the correct predicted proportion of the base classifier to the positive class sample, i.e., sensitivity, and TNR represents the correct predicted proportion of the base classifier to the negative class sample, i.e., specificity; />Representing the maximum G-mean value in all base classifiers; />Representing the minimum G-mean value in all base classifiers; wherein->，/>，/>Representing adaptive weights; at this time, the weight of each base classifier +.>The method comprises the following steps:
in the method, in the process of the invention,representing the k-th basis classifier adaptive comprehensive weight updating formula,/for>The self-adaptive comprehensive weight updating formula of the ith base classifier is represented, and N represents the number of the base learners;
voting results of ensemble learningThe method comprises the following steps:
in the method, in the process of the invention,representing the prediction result of the kth basis classifier.
The system is implemented by a cognitive assessment method based on virtual reality and integrated learning, and consists of a user subsystem, a virtual supermarket subsystem and a cognitive assessment subsystem, wherein:
the user subsystem is used for registering and storing personal information;
the virtual supermarket subsystem is used for collecting data of a user in the virtual supermarket;
the cognitive evaluation subsystem is used for evaluating the cognitive functions of the user, and the cognitive functions of the user are divided into normal cognitive and mild cognitive impairment.
Compared with the prior art, the invention has the following beneficial effects: (1) The method carries out balance treatment on a few samples, provides a KM-Borderlin eSMOTE algorithm, can effectively reduce model deviation and avoid over-fitting; (2) The method is based on relative accuracy, measurement of the performance index AUC of the classification model and unbalanced classification index G-mean, and provides a self-adaptive comprehensive weight updating integrated learning algorithm, so that the characteristics and rules of data are better captured, and the generalization capability and classification performance of the model are improved; (3) The system simulates a real shopping scene to carry out cognitive assessment on the elderly, and improves the ecological effectiveness of the cognitive assessment process.
Drawings
Fig. 1 is a schematic diagram of the steps of a cognitive assessment method based on virtual reality and ensemble learning according to the present invention.
Fig. 2 is a functional block diagram of a virtual reality and ensemble learning based cognitive assessment system of the present invention.
FIG. 3 is a diagram of the definition and description of the features of the data collected by the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following embodiments and the accompanying drawings, in order to make the objects, technical solutions and advantages of the present invention more apparent. The exemplary embodiments of the present invention and the descriptions thereof are used herein to explain the present invention, but are not intended to limit the invention.
The invention provides a cognitive assessment method based on virtual reality and ensemble learning, wherein the steps of the cognitive assessment method based on virtual reality and ensemble learning are shown in fig. 1, and the method comprises the following steps:
s100, dividing a final data set into a training set, a test set and a verification set according to the proportion of 8:1:1 after preprocessing the data;
s200, aiming at the problem of unbalanced quantity of two types of samples, providing a KM-Borderlin eSMOTE algorithm to up-sample a few types of samples;
s300, reducing the dimension of the features by using a principal component analysis algorithm, sorting the features in descending order according to the variance contribution rate after reducing the dimension, and taking the features with the accumulated variance contribution rate more than 95% as classification features;
s400, selecting an N-type base classifier for model training;
s500, integrating output results of the N-class base classifier by utilizing a self-adaptive comprehensive weight updating integrated learning algorithm to obtain a final classification class.
The KM-BorderlineSMOTE algorithm in the S200 comprises the following steps:
s210, calculating K neighbors of all minority class samples by using K neighbor algorithms, and dividing minority class sample properties at the same time: k neighbors of the ith minority class sample containMinority class samples and->Multiple types of samples, if->The properties of the minority class samples are classified as Safe; if->The properties of the minority class samples are divided into Danger; if->The properties of the minority class samples are classified as Noise;
s220, synthesizing a new sample:
s221, synthesizing new samples by using a Borderline-SMOTE algorithm on minority samples with Danger properties, wherein the number of required manual generation is as follows:
in the method, in the process of the invention,representing the number of minority class samples of the nature Danger, -/->The number of the minority class samples with the property of Safe is represented, and M represents the total number of the samples in the data set;
s222, for minority samples with Safe properties, firstly, clustering the minority samples with Safe properties by using an improved K-Medoids algorithm, and dividing a data set into K clusters; then setting sampling multiplying powerThe method comprises the following steps:
in the method, in the process of the invention,sample rate for the i-th cluster, +.>Represents the cluster center of the ith cluster, +.>Represents the j-th sample in the i-th cluster,/>Representation sample->To cluster center->K represents the number of clusters, +.>Representing the number of small samples in the ith cluster;
s223, synthesizing a new sample according to the sampling multiplying power:
in the method, in the process of the invention,representing a new sample of the composition, +.>And->Representing a minority class of samples and randomly selected samples in its K-nearest neighbor;
s230, combining the synthesized new sample with the original sample data set to obtain a final data set.
Wherein, the modified K-Medoids algorithm in S222 includes: the clustering criterion function is introduced to the K-Medoids algorithm, so that the influence of abnormal values is reduced, meanwhile, the dividing result is closer to the cluster center point, and the improved optimization objective function is as follows:
in the method, in the process of the invention,expressed as a change to the clustering criterion function, M represents the total number of samples in the dataset, <>Represents standard deviation of kth cluster, +.>Represents the cluster center of the kth cluster, +.>Represents the j-th sample in the k-th cluster,/>Representation sample->To cluster center->Is a distance of (3).
The self-adaptive comprehensive weight updating integrated learning algorithm in S500 includes: self-adaptive comprehensive weight updating formula based on relative accuracy, measuring performance index AUC of classification model and unbalanced classification index G-mean：
In the method, in the process of the invention,representing an adaptive comprehensive weight updating formula; />And->Respectively representing the maximum error rate and the minimum error rate in all the base classifiers; />Representing the error rate of the kth base classifier; />The area not covered by the ROC curve representing the kth basis classifier, i.e. +.>Wherein->Representing the area covered under the ROC curve of the kth basis classifier; />Representing the maximum of the uncovered areas of all the basis classifier ROC curves; />Representing the minimum of uncovered areas of all base classifier ROC curves; />G-mean value representing kth basis classifier, i.e.>Where TPR represents the correct predicted proportion of the base classifier to the positive class sample, i.e., sensitivity, and TNR represents the correct predicted proportion of the base classifier to the negative class sample, i.e., specificity; />Representing the maximum G-mean value in all base classifiers; />Representing the minimum G-mean value in all base classifiers; wherein->，/>，/>Representing adaptive weights; at this time, the weight of each base classifier +.>The method comprises the following steps:
in the method, in the process of the invention,representing the k-th basis classifier adaptive comprehensive weight updating formula,/for>The self-adaptive comprehensive weight updating formula of the ith base classifier is represented, and N represents the number of the base learners;
voting results of ensemble learningThe method comprises the following steps:
in the method, in the process of the invention,representing the prediction result of the kth basis classifier.
The invention provides a cognitive evaluation system based on virtual reality and integrated learning, which is composed of a user subsystem, a virtual supermarket subsystem and a cognitive evaluation subsystem, wherein a functional module diagram of the cognitive evaluation system based on virtual reality and integrated learning is shown in fig. 2, and the cognitive evaluation system comprises the following components:
the user subsystem is used for registering and storing personal information;
the virtual supermarket subsystem is used for collecting data of a user in the virtual supermarket;
the cognitive evaluation subsystem is used for evaluating the cognitive functions of the user, and the cognitive functions of the user are divided into normal cognitive and mild cognitive impairment.
The data preprocessing in S100 includes digitizing all data collected from the virtual supermarket subsystem, where the collected data includes user behavior data, cognitive dimension test game data, track data, and region data: the user behavior data has 11 characteristics, including the conditions of 5 types of actually purchased commodities on a shopping list, the quantity of the types of the purchased wrong commodities, the quantity of the wrong commodities, the correct payment amount, the correct input of payment passwords and the time of memorizing the detail of the commodities on a screen in two rounds; the cognitive dimension test game data has 4 characteristics, including the response speed performance in the naming game test data and the processing speed game test data, the correct number in the psychology movement capability game test data and the abstract capability game test data; the track data has 5 characteristics, including total moving distance, total time, average speed, normalized speed and speed standard deviation; zone data total 22 features including time and number of entries of participants in 11 zones: the collected data characteristic definition and description diagram is shown in fig. 3:
in particular, in the user behavior data, the i-th type of commodity actually purchased isThe definition is as follows:
in the method, in the process of the invention,，/>representing the number of items i actually purchased, +.>Representing the preset number of commodities i to be purchased;
reaction speed performance of ith game in cognitive dimension test game dataThe definition is as follows:
in the method, in the process of the invention,indicating the correct number in the ith game, +.>Indicating the j-th correct reaction rate in the i-th game;
average velocity in trace dataNormalized speed->Standard deviation of speed->The formula is as follows:
in the method, in the process of the invention,，/>，/>respectively representing the speed, the maximum speed and the minimum speed in the game;
wherein, the correct payment amount and the correct input payment password in the user behavior data use the thought of single-heat coding, and 0 and 1 are used for representing the characteristics; in order to eliminate the dimensional influence of different features and accelerate the convergence speed of the model, unified normalization processing is carried out on other data in the user behavior data, the cognitive dimension test game data, the track data and the region data.
Finally, what should be said is: the above embodiments are only for illustrating the technical aspects of the present invention, and although the present invention has been described in detail with reference to the above embodiments, it should be understood by those skilled in the art that: modifications and equivalents may be made thereto without departing from the spirit and scope of the invention, which is intended to be encompassed by the claims.
Claims (3)
1. The cognitive assessment method based on virtual reality and ensemble learning is characterized by comprising the following steps of:
s100, after preprocessing data, dividing a final data set into a training set, a test set and a verification set according to the proportion of 8:1:1, wherein the data comprises user behavior data, cognitive dimension test game data, track data and region data;
s200, aiming at the problem of unbalanced quantity of two types of samples, a KM-Borderlin eSMOTE algorithm is provided for upsampling a few types of samples:
s210, calculating K neighbors of all minority class samples by using K neighbor algorithms, and dividing minority class sample properties at the same time: k neighbors of the ith minority class sample containMinority class samples and->Multiple types of samples, if->The properties of the minority class samples are classified as Safe; if->The properties of the minority class samples are divided into Danger; if it isThe properties of the minority class samples are classified as Noise;
s220, synthesizing a new sample:
s221, synthesizing new samples by using a Borderline-SMOTE algorithm on minority samples with Danger properties, wherein the number of required manual generation is as follows:
in the method, in the process of the invention,representing the number of minority class samples of the nature Danger, -/->The number of the minority class samples with the property of Safe is represented, and M represents the total number of the samples in the data set;
s222, firstly, introducing a clustering criterion function on a K-Medoids algorithm, so that the influence of abnormal values is reduced, meanwhile, the dividing result is closer to a cluster center point, and the improved optimization objective function is as follows:
in the method, in the process of the invention,expressed as a change to the clustering criterion function, M represents the total number of samples in the dataset, <>Represents standard deviation of kth cluster, +.>Represents the cluster center of the kth cluster, +.>Represents the j-th sample in the k-th cluster,/>Representation sample->To the cluster centerIs a distance of (2);
clustering the minority samples with Safe properties by using an improved K-Medoids algorithm, and dividing a data set into K clusters; then setting sampling multiplying powerThe method comprises the following steps:
in the method, in the process of the invention,sample rate for the i-th cluster, +.>Clusters representing the ith clusterCenter (S)/(S)>Represents the j-th sample in the i-th cluster,/>Representation sample->To cluster center->K represents the number of clusters, +.>Representing the number of small samples in the ith cluster;
s223, synthesizing a new sample according to the sampling multiplying power:
in the method, in the process of the invention,representing a new sample of the composition, +.>And->Representing a minority class of samples and randomly selected samples in its K-nearest neighbor;
s230, merging the synthesized new sample with the original sample data set to obtain a final data set;
s300, reducing the dimension of the features by using a principal component analysis algorithm, sorting the features in descending order according to the variance contribution rate after reducing the dimension, and taking the features with the accumulated variance contribution rate more than 95% as classification features;
s400, selecting an N-type base classifier for model training;
and S500, integrating output results of the N-class base classifier by utilizing a self-adaptive comprehensive weight updating integrated learning algorithm to obtain final classification categories, namely normal cognition and mild cognition disorder.
2. The cognitive assessment method based on virtual reality and ensemble learning of claim 1, wherein the adaptive ensemble learning algorithm of updating the ensemble learning of weights in S500 includes: self-adaptive comprehensive weight updating formula based on relative accuracy, measuring performance index AUC of classification model and unbalanced classification index G-mean：
In the method, in the process of the invention,representing an adaptive comprehensive weight updating formula; />And->Respectively representing the maximum error rate and the minimum error rate in all the base classifiers; />Representing the error rate of the kth base classifier; />The area not covered by the ROC curve representing the kth basis classifier, i.e. +.>Wherein->Representing the area covered under the ROC curve of the kth basis classifier; />Representing the maximum of the uncovered areas of all the basis classifier ROC curves; />Representing the minimum of uncovered areas of all base classifier ROC curves; />G-mean value representing kth basis classifier, i.e.>Where TPR represents the correct predicted proportion of the base classifier to the positive class sample, i.e., sensitivity, and TNR represents the correct predicted proportion of the base classifier to the negative class sample, i.e., specificity; />Representing the maximum G-mean value in all base classifiers; />Representing the minimum G-mean value in all base classifiers; wherein the method comprises the steps of，/>，/>Representing adaptive weights; at this time, the weight of each base classifier +.>The method comprises the following steps:
in the method, in the process of the invention,representing the k-th basis classifier adaptive comprehensive weight updating formula,/for>The self-adaptive comprehensive weight updating formula of the ith base classifier is represented, and N represents the number of the base learners;
voting results of ensemble learningThe method comprises the following steps:
in the method, in the process of the invention,representing the prediction result of the kth basis classifier.
3. A cognitive assessment system based on virtual reality and ensemble learning, the system being implemented by the method of any of claims 1-2, characterized in that the system consists of a user subsystem, a virtual supermarket subsystem, a cognitive assessment subsystem:
the user subsystem is used for registering and storing personal information;
the virtual supermarket subsystem is used for collecting data of a user in the virtual supermarket;
the cognitive evaluation subsystem is used for evaluating the cognitive functions of the user, and the cognitive functions of the user are divided into normal cognitive and mild cognitive impairment.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410122491.0A CN117648572B (en) | 2024-01-30 | 2024-01-30 | Cognitive assessment method and system based on virtual reality and ensemble learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410122491.0A CN117648572B (en) | 2024-01-30 | 2024-01-30 | Cognitive assessment method and system based on virtual reality and ensemble learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117648572A CN117648572A (en) | 2024-03-05 |
| CN117648572B true CN117648572B (en) | 2024-04-05 |
Family
ID=90046415
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410122491.0A Active CN117648572B (en) | 2024-01-30 | 2024-01-30 | Cognitive assessment method and system based on virtual reality and ensemble learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117648572B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111466876A (en) * | 2020-03-24 | 2020-07-31 | 山东大学 | Alzheimer's disease auxiliary diagnosis system based on fNIRS and graph neural network |
| CN111681132A (en) * | 2020-06-17 | 2020-09-18 | 四川大学 | Typical power consumption mode extraction method suitable for massive class unbalanced load data |
| CN113842144A (en) * | 2021-09-26 | 2021-12-28 | 常州大学 | Self-adaptive multiband electroencephalogram signal mild cognitive impairment evaluation method |
| CN115601604A (en) * | 2022-11-29 | 2023-01-13 | 西南石油大学（Cn） | Multi-task microbubble track tracking method based on long-time and short-time memory network |
| CN115691804A (en) * | 2022-10-19 | 2023-02-03 | 深圳脑吾脑网络科技有限公司 | Social cognition assessment training system based on digital twin-neuron universe |
| CN116092673A (en) * | 2023-04-10 | 2023-05-09 | 华南理工大学 | Portable multi-information fusion analysis and intervention evaluation system and method thereof |
| CN117219127A (en) * | 2023-08-30 | 2023-12-12 | 腾讯科技（深圳）有限公司 | Cognitive state recognition method and related equipment |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20190274610A1 (en) * | 2018-03-09 | 2019-09-12 | Yongbo Wang | Cognitive and Neuropsychic Improvement and Evaluation/Analysis of Alzheimer's Disease and Other Neurological Diseases with Augmented and Mixed Reality |
| JP7491943B2 (en) * | 2019-03-22 | 2024-05-28 | コグノア，インク． | Personalized digital therapeutic method and device |
| US20220293266A1 (en) * | 2021-03-09 | 2022-09-15 | The Hong Kong Polytechnic University | System and method for detection of impairment in cognitive function |
-
2024
- 2024-01-30 CN CN202410122491.0A patent/CN117648572B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111466876A (en) * | 2020-03-24 | 2020-07-31 | 山东大学 | Alzheimer's disease auxiliary diagnosis system based on fNIRS and graph neural network |
| CN111681132A (en) * | 2020-06-17 | 2020-09-18 | 四川大学 | Typical power consumption mode extraction method suitable for massive class unbalanced load data |
| CN113842144A (en) * | 2021-09-26 | 2021-12-28 | 常州大学 | Self-adaptive multiband electroencephalogram signal mild cognitive impairment evaluation method |
| CN115691804A (en) * | 2022-10-19 | 2023-02-03 | 深圳脑吾脑网络科技有限公司 | Social cognition assessment training system based on digital twin-neuron universe |
| CN115601604A (en) * | 2022-11-29 | 2023-01-13 | 西南石油大学（Cn） | Multi-task microbubble track tracking method based on long-time and short-time memory network |
| CN116092673A (en) * | 2023-04-10 | 2023-05-09 | 华南理工大学 | Portable multi-information fusion analysis and intervention evaluation system and method thereof |
| CN117219127A (en) * | 2023-08-30 | 2023-12-12 | 腾讯科技（深圳）有限公司 | Cognitive state recognition method and related equipment |
Non-Patent Citations (5)
| Title |
|---|
| "Novel K-Medoids Based SMOTE Integrated With Locality Preserving Projections for Fault Diagnosis";Q. -X. Zhu等;《IEEE Transactions on Instrumentation and Measurement》;20221101;第71卷;第1-8页 * |
| "基于大脑结构形态学的老化模式分析与脑龄预测研究";李姗;《中国优秀硕士学位论文全文数据库 医药卫生科技辑》;20230115(第1期);E080-2 * |
| "Bio-inspired dimensionality reduction for Parkinson’s disease (PD) classification";Pasha, A.等;《Health Information Science and Systems》;20200309;第8卷;第1-22页 * |
| "基于K-medoids-NCA-SMOTE-BSVM融合模型的网络交易平台高质量数据资源识别研究";倪渊等;《运筹与管理》;20231125;第32卷(第11期);第87-93页 * |
| Sarica, A.等."Introducing the Rank-Biased Overlap as Similarity Measure for Feature Importance in Explainable Machine Learning: A Case Study on Parkinson's Disease".《International Conference on Brain Informatics Springer》.2022,第129–139页. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117648572A (en) | 2024-03-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Çağlar et al. | Automatic recognition of Parkinson’s disease from sustained phonation tests using ANN and adaptive neuro-fuzzy classifier | |
| CN103714261B (en) | Intelligent auxiliary medical treatment decision supporting method of two-stage mixed model | |
| Liu et al. | Multi-LSTM networks for accurate classification of attention deficit hyperactivity disorder from resting-state fMRI data | |
| Herath et al. | Autism spectrum disorder diagnosis support model using Inception V3 | |
| Xia et al. | Dynamic viewing pattern analysis: towards large-scale screening of children with ASD in remote areas | |
| Ang et al. | Explored research on data preprocessing and mining technology for clinical data applications | |
| CN117648572B (en) | Cognitive assessment method and system based on virtual reality and ensemble learning | |
| CN114155306A (en) | Visual analysis method for fundus diseases through computer vision and deep learning | |
| CN117219127A (en) | Cognitive state recognition method and related equipment | |
| Abhishek et al. | Diagnosis of Parkinson’s disorder through speech data using machine learning algorithms | |
| CN116484290A (en) | Depression recognition model construction method based on Stacking integration | |
| Alattab et al. | A smart diseases diagnosis and classification strategy of electronic healthcare application using novel hybrid artificial intelligence approaches | |
| Yildirim et al. | A new hybrid approach based on AOA, CNN and feature fusion that can automatically diagnose Parkinson's disease from sound signals: PDD-AOA-CNN | |
| Umar et al. | Classification Using Radial Basis Function for Prediction of Parkinson's Disease | |
| Zhang et al. | Application of the rough set theory and BP neural network model in disease diagnosis | |
| Panda et al. | Machine Learning-Based Framework for Early Detection of Distinguishing Different Stages of Parkinson's Disease. | |
| CN108346471A (en) | A kind of analysis method and device of pathological data | |
| Wijaya et al. | Gradient Boosted Tree Based Feature Selection and Parkinson’s Disease Classification | |
| Wisler et al. | Can machines learn continuous measures of speech severity from ordinal training labels? | |
| Fathima et al. | Analysis Of Machine Learning Algorithm for Prognostication of Parkinson’s Disease Using Speech Articulation | |
| Surya Narayana et al. | Predicting students’ transformation to maximum depressive disorder and level of suicidal tendency | |
| Kumar et al. | Early Prediction of Parkinson’s Disease using Multiple SVM Classifiers | |
| Karthikaa et al. | Early Diagnosis of Parkinson’s Disease Using Machine Learning | |
| CN112201355B (en) | Construction method of health evaluation iterative classifier model | |
| Chatterjee et al. | Machine learning and deep learning-based advanced classification techniques for the detection of major depressive disorder |
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 |