Personal cognitive impairment risk assessment method and apparatus, terminal, and storage medium

Abstract

A personal cognitive impairment risk assessment method and apparatus, a terminal, and a storage medium. The method comprises: acquiring target data, wherein the target data includes test data, feature data, and medical record data of a target patient; and inputting the target data into a trained cognitive impairment risk assessment model, and processing the target data by means of the cognitive impairment risk assessment model to obtain a cognitive impairment risk assessment result of the target patient. The trained cognitive impairment risk assessment model is used to process the target data to obtain the cognitive impairment risk assessment result of the target patient. The efficiency of personal cognitive impairment risk assessment is effectively improved.

Description

A personal cognitive impairment risk assessment method, device, terminal and storage medium TECHNICAL FIELD

[0001] The present application relates to the field of artificial intelligence medical technology, and in particular to a personal cognitive impairment risk assessment method, device, terminal and storage medium. BACKGROUND

[0002] Traditional cognitive impairment assessment methods usually rely on evaluators to assess subjects item by item, such as memory, attention, language ability, executive function, etc., and finally submit the assessment results to doctors for comprehensive judgment. However, this traditional assessment method highly depends on manual assessment based on multiple test results, resulting in low efficiency of personal cognitive impairment risk assessment.

[0003] Therefore, the prior art has defects and needs to be improved and developed.

[0004] SUMMARY

[0005] The technical problem to be solved by the present application is to provide a personal cognitive impairment risk assessment method, device, terminal and storage medium to solve the problem of low efficiency of personal cognitive impairment risk assessment in the prior art.

[0006] The technical solution adopted by the present application to solve the technical problem is as follows:

[0007] In a first aspect, the present application provides a personal cognitive impairment risk assessment method, which comprises:

[0008] Obtaining target data, the target data being detection data, feature data and medical record data of a target patient;

[0009] Inputting the target data into a trained cognitive impairment risk assessment model, and obtaining a cognitive impairment risk assessment result of the target patient by processing the cognitive impairment risk assessment model.

[0010] In an embodiment, the step of inputting the target data into the trained cognitive impairment risk assessment model and processing the cognitive impairment risk assessment model to obtain the cognitive impairment risk assessment result of the target patient comprises:

[0011] Inputting the detection data and the medical record data into a feature extraction module of the cognitive impairment risk assessment model, and obtaining a plurality of original features by processing the feature extraction module;

[0012] Inputting all the original features into a feature perception module of the cognitive impairment risk assessment model, and obtaining a node embedding vector of each original feature by processing the feature perception module;

[0013] input the node embedding vectors of all the original features into a fusion module of the cognitive impairment risk assessment model, and obtain a spliced node embedding vector through processing by the fusion module;

[0014] input the spliced node embedding vector and the feature data into a risk assessment module of the cognitive impairment risk assessment model, and obtain a cognitive impairment risk assessment result of the target patient through processing by the risk assessment module.

[0015] In an implementation, the inputting of the node embedding vectors of all the original features into the fusion module of the cognitive impairment risk assessment model and the obtaining of the spliced node embedding vector through processing by the fusion module include:

[0016] inputting the node embedding vectors of all the original features into the fusion module of the cognitive impairment risk assessment model;

[0017] performing average aggregation processing or splicing processing on all the node embedding vectors by using the fusion module to obtain the spliced node embedding vector.

[0018] In an implementation, the inputting of the spliced node embedding vector and the feature data into the risk assessment module of the cognitive impairment risk assessment model and the obtaining of the cognitive impairment risk assessment result of the target patient through processing by the risk assessment module include:

[0019] inputting the spliced node embedding vector and the feature data into the risk assessment module of the cognitive impairment risk assessment model, performing splicing processing on the spliced node embedding vector and the feature data by using a linear transformation layer in the risk assessment module to obtain a fusion embedding vector;

[0020] inputting the fusion embedding vector into a classifier of the risk assessment module, and obtaining classification information of the target patient through processing by the classifier;

[0021] obtaining the cognitive impairment risk assessment result of the target patient based on the classification information of the target patient.

[0022] In an implementation, the classifier is a multilayer perceptron; and the inputting of the fusion embedding vector into the classifier of the risk assessment module and the obtaining of the classification information of the target patient through processing by the classifier include:

[0023] inputting the fusion embedding vector into an input layer of the multilayer perceptron, and transmitting the fusion embedding vector to an intermediate hidden layer of the multilayer perceptron through the input layer;

[0024] transforming the fusion embedding vector by using each neuron in the intermediate hidden layer to perform a nonlinear transformation to obtain intermediate hidden layer output data, and transmitting the intermediate hidden layer output data to an output layer of the multilayer perceptron;

[0025] processing the intermediate hidden layer output data by using the output layer to obtain classification information of the target patient, the classification information being a cognitive impairment category and a probability distribution thereof.

[0026] In an implementation, the cognitive impairment risk assessment result of the target patient based on the classification information of the target patient comprises:

[0027] screening the cognitive impairment category with the highest probability in the classification information of the target patient;

[0028] taking the screened cognitive impairment category with the highest probability as the cognitive impairment risk assessment result of the target patient.

[0029] In an implementation, the training process of the cognitive impairment risk assessment model comprises:

[0030] constructing a training data set;

[0031] constructing an initial cognitive impairment risk assessment model;

[0032] inputting the training data set into the initial cognitive impairment risk assessment model to be trained, processing the training data set by using the initial cognitive impairment risk assessment model to obtain an initial cognitive impairment risk assessment result;

[0033] calculating a loss function based on the initial cognitive impairment risk assessment result and a preset label, minimizing the loss function by using an optimization algorithm, and updating model parameters;

[0034] iteratively updating the model parameters until an end condition is reached to obtain a trained cognitive impairment risk assessment model.

[0035] In a second aspect, an embodiment of the present application further provides a personal cognitive impairment risk assessment device, comprising:

[0036] a data acquisition module configured to acquire target data, the target data being detection data, feature data, and medical record data of a target patient;

[0037] a risk assessment module configured to input the target data into a trained cognitive impairment risk assessment model, process the target data by using the cognitive impairment risk assessment model, and obtain a cognitive impairment risk assessment result of the target patient.

[0038] In a third aspect, an embodiment of the present application further provides a terminal, comprising a memory, a processor, and a personal cognitive impairment risk assessment program stored in the memory and executable on the processor, wherein the personal cognitive impairment risk assessment program, when executed by the processor, implements the steps of the personal cognitive impairment risk assessment method as described above.

[0039] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium storing a personal cognitive impairment risk assessment program executable to implement the steps of the personal cognitive impairment risk assessment method as described above.

[0040] The present application has the following beneficial effects: The embodiment of the present application obtains target data, which is detection data, feature data, and medical record data of a target patient; inputs the target data into a trained cognitive impairment risk assessment model, and obtains a cognitive impairment risk assessment result of the target patient through processing by the cognitive impairment risk assessment model. The present application effectively improves the efficiency of personal cognitive impairment risk assessment by processing target data using a trained cognitive impairment risk assessment model to obtain a cognitive impairment risk assessment result of a target patient. BRIEF DESCRIPTION OF DRAWINGS

[0041] FIG. 1 is a flowchart of a preferred embodiment of the personal cognitive impairment risk assessment method of the present application.

[0042] FIG. 2 is a structural schematic diagram of a preferred embodiment of the personal cognitive impairment risk assessment device of the present application.

[0043] FIG. 3 is a principle block diagram of the terminal of the present application. DETAILED DESCRIPTION

[0044] To make the objectives, technical solutions, and advantages of the present application clearer and more explicit, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and do not limit the present application.

[0045] The traditional cognitive impairment assessment method usually relies on an evaluator to perform item-by-item assessment on a subject, such as memory, attention, language ability, and executive function, and finally submits the assessment results to a doctor for comprehensive judgment. However, this traditional assessment method highly depends on manual assessment based on multiple check item results, and the efficiency of manual information integration and assessment is low, and the efficiency of personal cognitive impairment risk assessment is low.

[0046] In view of the above defects of the prior art, the present application provides a personal cognitive impairment risk assessment method, device, terminal and storage medium, the method comprising: obtaining target data, the target data being detection data, feature data and medical record data of a target patient; inputting the target data into a trained cognitive impairment risk assessment model, and obtaining a cognitive impairment risk assessment result of the target patient through processing by the cognitive impairment risk assessment model. The present application effectively improves the efficiency of personal cognitive impairment risk assessment by processing the target data using the trained cognitive impairment risk assessment model to obtain the cognitive impairment risk assessment result of the target patient.

[0047] Referring to FIG. 1, the personal cognitive impairment risk assessment method according to the present application comprises the following steps:

[0048] In step S100, target data is obtained, the target data being detection data, feature data and medical record data of a target patient.

[0049] Specifically, the target patient is a patient who undergoes personal cognitive impairment risk assessment in a medical institution. The detection data includes memory test data, attention test data, executive ability test data, language ability test data, visual space test data and directional ability test data. The feature data includes age, place of origin, occupation and hobbies. The medical record data is medical record information of the patient diagnosed by a doctor. Since the target data includes detection data, feature data and medical record data, the condition of the target patient can be comprehensively and multi-dimensionally reflected, and these data complement each other to reduce the deviation caused by single data, thereby improving the accuracy of cognitive impairment risk assessment.

[0050] Referring to FIG. 1, the personal cognitive impairment risk assessment method according to the present application further comprises the following steps:

[0051] In step S200, the target data is input into a trained cognitive impairment risk assessment model, and a cognitive impairment risk assessment result of the target patient is obtained through processing by the cognitive impairment risk assessment model.

[0052] Specifically, when it is necessary to perform risk assessment on the target patient, the target data is input into the trained cognitive impairment risk assessment model to obtain the cognitive impairment risk assessment result corresponding to the target patient.

[0053] The traditional cognitive impairment risk assessment method usually relies on the experience and subjective judgment of doctors, which is easily affected by the personal experience, knowledge level and evaluation standard of doctors, resulting in subjectivity and inconsistency of the evaluation results. The cognitive impairment risk assessment model provided by the present application avoids the interference of subjective factors of human evaluation by comprehensively analyzing the detection data, feature data and medical record data of the target patient, and the evaluation results of the model are more objective and accurate, which can more truly reflect the cognitive impairment risk level of the target patient. In addition, the traditional cognitive impairment risk assessment method requires doctors to spend a lot of time and effort for manual evaluation, which is low in efficiency and difficult to meet the needs of large-scale population screening. The present application uses the trained cognitive impairment risk assessment model for automatic evaluation, which greatly improves the evaluation efficiency. In the scene of community elderly dementia risk screening, a large number of people can be quickly evaluated for risk, and high-risk groups can be found in time to provide a basis for subsequent intervention measures. The cognitive impairment risk assessment model provided by the present application overcomes the limitations of traditional evaluation methods and provides a more scientific, efficient and scalable solution for cognitive impairment risk assessment, which has important application value and social significance.

[0054] In one implementation, the target data is input into the trained cognitive impairment risk assessment model, and the cognitive impairment risk assessment result of the target patient is obtained by processing the cognitive impairment risk assessment model.

[0055] The detection data and the medical record data are input into the feature extraction module of the cognitive impairment risk assessment model, and a plurality of original features are obtained by processing the feature extraction module.

[0056] The node embedding vectors of all the original features are input into the feature perception module of the cognitive impairment risk assessment model, and the node embedding vectors of each original feature are obtained by processing the feature perception module.

[0057] The node embedding vectors of all the original features are input into the fusion module of the cognitive impairment risk assessment model, and the spliced node embedding vectors are obtained by processing the fusion module.

[0058] The spliced node embedding vectors and the feature data are input into the risk assessment module of the cognitive impairment risk assessment model, and the cognitive impairment risk assessment result of the target patient is obtained by processing the risk assessment module.

[0059] Specifically, the cognitive impairment risk assessment model includes a feature extraction module, a feature perception module, a fusion module and a risk assessment module. The present application can make up for the deficiency of single data information by fusing detection data, medical record data and feature data, and the result of cognitive impairment risk assessment can be more accurate because each kind of data contains potential information related to cognitive impairment risk.

[0060] In the feature-aware module processing the original features, the original features are regarded as nodes, the attention weight between each node and neighbor node is calculated, and then the node embedding vector of each original feature is obtained based on the attention weight between each node and neighbor node. These node embedding vectors can better capture the dependencies and relationships between the original features.

[0061] In an implementation manner, the fusion module inputs the node embedding vectors of all the original features into the cognitive impairment risk assessment model, and the fusion module processes to obtain a spliced node embedding vector, including:

[0062] The fusion module inputs the node embedding vectors of all the original features into the cognitive impairment risk assessment model;

[0063] The fusion module performs average aggregation processing or splicing processing on all the node embedding vectors to obtain a spliced node embedding vector.

[0064] Specifically, the original features are processed by average aggregation processing or splicing processing to obtain a spliced node embedding vector. Through the fusion process, redundant information between different original features can be removed, and more valuable information can be retained, so that the model pays more attention to the key features between data, reduces the sensitivity of the model to noise data, thereby improving the robustness and generalization ability of the model, and enabling the model to maintain good performance when facing unknown data.

[0065] In an implementation manner, the risk assessment module inputs the spliced node embedding vector and the feature data into the cognitive impairment risk assessment model, and the risk assessment module processes to obtain a cognitive impairment risk assessment result of a target patient, including:

[0066] The risk assessment module inputs the spliced node embedding vector and the feature data into the cognitive impairment risk assessment model, and a linear transformation layer in the risk assessment module performs splicing processing on the spliced node embedding vector and the feature data to obtain a fusion embedding vector;

[0067] The fusion embedding vector is input into a classifier in the risk assessment module, and the classifier processes to obtain classification information of the target patient;

[0068] Based on the classification information of the target patient, a cognitive impairment risk assessment result of the target patient is obtained.

[0069] Specifically, the risk assessment module comprises a linear transformation layer and a classifier. The linear transformation layer is used to align the dimensions of the spliced node embedding vector and the feature data, and obtain a fusion embedding vector. Then, the fusion embedding vector is input into the classifier to obtain classification information, and the cognitive impairment risk assessment result of the target patient is obtained based on the classification information. In this way, the cognitive impairment risk of the target patient can be comprehensively, efficiently and accurately evaluated, and strong support can be provided for personalized intervention, which is of great significance for cognitive impairment management and intervention.

[0070] In an implementation manner, the classifier is a multilayer perceptron; the fusion embedding vector is input into the input layer of the multilayer perceptron, and the fusion embedding vector is transmitted to the intermediate hidden layer of the multilayer perceptron through the input layer.

[0071] The fusion embedding vector is input into the input layer of the multilayer perceptron, and the fusion embedding vector is transmitted to the intermediate hidden layer of the multilayer perceptron through the input layer.

[0072] Each neuron in the intermediate hidden layer performs a non-linear transformation on the fusion embedding vector to obtain intermediate hidden layer output data, and the intermediate hidden layer output data is transmitted to the output layer of the multilayer perceptron.

[0073] The output layer processes the intermediate hidden layer output data to obtain the classification information of the target patient, and the classification information is the cognitive impairment category and its probability distribution.

[0074] Specifically, the classifier of the present application is a multilayer perceptron (MLP, Multilayer Perceptron), which is a forward structure artificial neural network and can map a set of input vectors to a set of output vectors. The multilayer perceptron can learn complex patterns and relationships in data, and can better process the non-linear information contained in the fusion embedding vector compared with the linear classifier, thereby improving the accuracy of classification. The classifier of the present application specifically comprises an input layer, an intermediate hidden layer and an output layer. The input layer receives data, the intermediate hidden layer extracts features, and the output layer outputs a prediction result. In the intermediate hidden layer, each node is a neuron with a non-linear activation function, and the activation function is a Relu activation function (i.e. a linear rectifier function), which can be expressed as Relu(x)=max(x,0). In the formula, x is the input of the neuron. The Relu activation function introduces a non-linear factor for the neuron, which enables the neural network to approximate any complex non-linear function, thereby solving more extensive problems. In the output layer, the number of neurons in the output layer is consistent with the number of preset cognitive impairment categories.

[0075] In an implementation manner, the cognitive impairment risk assessment result of the target patient is obtained based on the classification information of the target patient, which comprises:

[0076] screening the highest probability cognitive disorder category in the classification information of the target patient;

[0077] taking the screened highest probability cognitive disorder category as the cognitive disorder risk assessment result of the target patient.

[0078] Specifically, when the classification information is obtained, the highest probability cognitive disorder category is taken as the cognitive disorder risk assessment result of the target patient, which can encourage the patient and the medical team to take intervention measures as early as possible according to the cognitive disorder risk assessment result, so as to slow down the progression of the disease and improve the quality of life of the patient, and reduce the economic burden on the family and society. For example, if the classification information is that the probability of mild cognitive impairment is 70%, the probability of Alzheimer's disease is 20%, and the probability of vascular dementia is 10%, then mild cognitive impairment is taken as the cognitive disorder risk assessment result of the patient.

[0079] In an embodiment, the training process of the cognitive disorder risk assessment model comprises:

[0080] constructing a training data set;

[0081] constructing an initial cognitive disorder risk assessment model;

[0082] inputting the training data set into the initial cognitive disorder risk assessment model to be trained, processing the training data set through the initial cognitive disorder risk assessment model, and obtaining an initial cognitive disorder risk assessment result;

[0083] calculating a loss function based on the initial cognitive disorder risk assessment result and a preset label, minimizing the loss function through an optimization algorithm, and updating the model parameters;

[0084] iteratively updating the model parameters until a termination condition is reached, and obtaining a trained cognitive disorder risk assessment model.

[0085] Specifically, the detection data, medical record data and feature data of the elderly with cognitive disorders for more than 5 years are obtained from multiple data sources as initial data. The elderly are people over 60 years old. Multiple data sources refer to different medical institutions, hospitals, families and communities, etc. After obtaining the initial data, data cleaning is needed because the initial data comes from multiple data sources. Data cleaning includes removing redundant data and removing abnormal data. For example, when there are duplicate records of the same patient in different data sources, redundant data needs to be removed. When removing abnormal data, the algorithm based on the k-means clustering idea can be used to find outliers and determine abnormal data. After the initial data is cleaned, the training data set is obtained. Since the training data set has the characteristics of multi-source, multi-dimension and multi-modal, it can help the model to better learn the data features and rules, thereby improving the prediction accuracy and generalization ability of the model.

[0086] Since the data in the training data set belongs to time series type data, the data contains time sequence information, and the historical data can potentially affect the future data. Therefore, in the initial cognitive impairment risk assessment model, it is considered that there is a potential graph structure, and each examination data, medical record data and feature data of the elderly are regarded as a node in the potential graph. Through the input data, information transmission is performed on other nodes in the potential graph to capture the potential time-varying dependency between different data. In the training process, the influence factors between different neighbor nodes are extracted based on the extracted information, and the relevance of different nodes is learned, the importance of the nodes in the time series at different time points is calculated, the overall information perception is performed, and finally the perceived information is fused to obtain a fusion embedding vector. The fusion embedding vector is input into the classifier, and the classifier uses the supervised learning method of the back propagation algorithm to train. Based on the initial cognitive impairment risk assessment result and the preset label, a loss function is calculated, the loss function is minimized through an optimization algorithm, and the model parameters are updated; the model parameters are iteratively updated until the end condition is reached, and a trained cognitive impairment risk assessment model is obtained. The end condition is that the iteration reaches a preset iteration number or the loss function value converges below a preset threshold. The present application can obtain a cognitive impairment risk assessment model with good performance through multiple rounds of iterative training of the model by introducing a multi-source, multi-dimensional and multi-modal feature training data set, thereby effectively ensuring the accuracy of the cognitive impairment risk assessment result.

[0087] In summary, the present application obtains target data, which is detection data, feature data and medical record data of a target patient, inputs the target data into a trained cognitive impairment risk assessment model, and obtains a cognitive impairment risk assessment result of the target patient through the cognitive impairment risk assessment model. The present application processes the target data by using the trained cognitive impairment risk assessment model to obtain the cognitive impairment risk assessment result of the target patient, thereby effectively improving the efficiency of individual cognitive impairment risk assessment in large-scale screening.

[0088] In one embodiment, as shown in FIG. 2, based on the above-mentioned individual cognitive impairment risk assessment method, the present application also correspondingly provides an individual cognitive impairment risk assessment device, which comprises:

[0089] A data acquisition module 100 is configured to acquire target data, wherein the target data is detection data, feature data and medical record data of a target patient;

[0090] A risk assessment module 200 is configured to input the target data into a trained cognitive impairment risk assessment model, and obtain a cognitive impairment risk assessment result of the target patient through the cognitive impairment risk assessment model.

[0091] In an implementation manner, the risk assessment module comprises:

[0092] a data input unit configured to input the detection data and the medical record data into a feature extraction module of the cognitive impairment risk assessment model, and obtain a plurality of original features through processing of the feature extraction module;

[0093] a feature perception unit configured to input all the original features into a feature perception module of the cognitive impairment risk assessment model, and obtain a node embedding vector of each of the original features through processing of the feature perception module;

[0094] a fusion processing unit configured to input the node embedding vectors of all the original features into a fusion module of the cognitive impairment risk assessment model, and obtain a spliced node embedding vector through processing of the fusion module;

[0095] an assessment result generation unit configured to input the spliced node embedding vector and the feature data into a risk assessment module of the cognitive impairment risk assessment model, and obtain a cognitive impairment risk assessment result of the target patient through processing of the risk assessment module.

[0096] In an implementation manner, the fusion processing unit comprises:

[0097] a first feature input unit configured to input the node embedding vectors of all the original features into the fusion module of the cognitive impairment risk assessment model;

[0098] a spliced node embedding vector generation unit configured to perform average aggregation processing or splicing processing on all the node embedding vectors by using the fusion module, and obtain a spliced node embedding vector.

[0099] In an implementation manner, the assessment result generation unit comprises:

[0100] a second feature input unit configured to input the spliced node embedding vector and the feature data into the risk assessment module of the cognitive impairment risk assessment model, perform splicing processing on the spliced node embedding vector and the feature data by using a linear transformation layer in the risk assessment module, and obtain a fusion embedding vector;

[0101] a classification information generation unit configured to input the fusion embedding vector into a classifier of the risk assessment module, and obtain classification information of the target patient through processing of the classifier;

[0102] a result generation unit configured to obtain the cognitive impairment risk assessment result of the target patient based on the classification information of the target patient.

[0103] In an implementation manner, the classifier is a multilayer perceptron; and the classification information generation unit comprises:

[0104] a fusion embedding vector input unit configured to input the fusion embedding vector to an input layer of the multi-layer perceptron, and transmit the fusion embedding vector to an intermediate hidden layer of the multi-layer perceptron via the input layer;

[0105] an intermediate hidden layer processing unit configured to perform a nonlinear transformation on the fusion embedding vector by using each neuron in the intermediate hidden layer to obtain intermediate hidden layer output data, and transmit the intermediate hidden layer output data to an output layer of the multi-layer perceptron;

[0106] an output layer processing unit configured to process the intermediate hidden layer output data by using the output layer to obtain classification information of the target patient, the classification information being a cognitive impairment category and a probability distribution thereof.

[0107] In an implementation manner, the result generation unit comprises:

[0108] a category screening unit configured to screen the cognitive impairment category with the highest probability in the classification information of the target patient;

[0109] a result determination unit configured to determine the cognitive impairment category with the highest probability screened as a cognitive impairment risk assessment result of the target patient.

[0110] In an implementation manner, the apparatus further comprises:

[0111] a data set construction unit configured to construct a training data set;

[0112] a model construction unit configured to construct an initial cognitive impairment risk assessment model;

[0113] a data set input unit configured to input the training data set into the initial cognitive impairment risk assessment model to be trained, and process the initial cognitive impairment risk assessment model to obtain an initial cognitive impairment risk assessment result;

[0114] a model parameter updating unit configured to calculate a loss function based on the initial cognitive impairment risk assessment result and a preset label, minimize the loss function by using an optimization algorithm, and update model parameters;

[0115] a model training unit configured to iteratively update the model parameters until an end condition is reached to obtain a trained cognitive impairment risk assessment model.

[0116] Based on the above-mentioned embodiments, the present application further provides a terminal, a principle block diagram of which can be shown in FIG. 3. The terminal comprises a processor, a memory, a network interface and a display screen connected through a system bus. The processor of the terminal is configured to provide computing and control capabilities. The memory of the terminal comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a personal cognitive impairment risk assessment device. The internal memory provides an environment for the operating system and the personal cognitive impairment risk assessment device in the non-volatile storage medium. The network interface of the terminal is configured to communicate with external terminals through a network connection. The personal cognitive impairment risk assessment device, when executed by the processor, implements the steps of any one of the personal cognitive impairment risk assessment methods. The display screen of the terminal can be a liquid crystal display screen or an electronic ink display screen.

[0117] Those skilled in the art can understand that the principle block diagram shown in FIG. 3 is only a block diagram of part of the structure related to the present application, and does not constitute a limitation on the terminal to which the present application is applied. The specific terminal can comprise more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement.

[0118] In one embodiment, a terminal is provided, which comprises a memory, a processor and a personal cognitive impairment risk assessment program stored in the memory and executable on the processor. The personal cognitive impairment risk assessment program, when executed by the processor, implements the steps of any one of the personal cognitive impairment risk assessment methods provided by the present application.

[0119] The present application further provides a computer readable storage medium, which stores a personal cognitive impairment risk assessment program. The personal cognitive impairment risk assessment program, when executed by a processor, implements the steps of any one of the personal cognitive impairment risk assessment methods provided by the present application.

[0120] It should be understood that the sequence of the steps in the above-mentioned embodiments does not mean the order of execution. The execution order of the processes should be determined according to their functions and inherent logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0121] Those skilled in the art can clearly understand that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the above-mentioned device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can be physically present separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be realized in the form of hardware or software function unit. In addition, the specific name of each functional unit and module is only for easy distinction, and does not limit the protection scope of the present application. The specific working process of the unit and module in the above-mentioned system can refer to the corresponding process in the foregoing method embodiment, which will not be described here.

[0122] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the related description of other embodiments.

[0123] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or a combination of computer software and electronic hardware. Whether the functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals can use different ways to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0124] In the embodiments provided by the present application, it should be understood that the disclosed device / terminal equipment and method can be implemented in other ways. For example, the above-mentioned device / terminal equipment embodiments are only schematic, for example, the division of the above-mentioned modules or units is only a logical function division, and actual implementation can be in another division manner, for example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed.

[0125] The above-mentioned embodiments are only used to illustrate the technical solutions of the present application, but not limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand; the technical solutions recorded in the foregoing embodiments can be modified, or some technical features can be replaced by equivalents; and these modifications or replacements do not deviate from the spirit and scope of the corresponding technical solutions, and should be included in the protection scope of the present application.

Claims

1. A method of personal cognitive impairment risk assessment, characterized by, The method comprises: obtaining target data, the target data being detection data, feature data and medical record data of a target patient; inputting the target data into a trained cognitive impairment risk assessment model, and obtaining a cognitive impairment risk assessment result of the target patient through processing by the cognitive impairment risk assessment model.

2. The method of claim 1, wherein, The method comprises: inputting the detection data and the medical record data into a feature extraction module of the cognitive impairment risk assessment model, and obtaining a plurality of original features through processing by the feature extraction module; inputting all the original features into a feature perception module of the cognitive impairment risk assessment model, and obtaining a node embedding vector of each original feature through processing by the feature perception module; inputting the node embedding vectors of all the original features into a fusion module of the cognitive impairment risk assessment model, and obtaining a spliced node embedding vector through processing by the fusion module; inputting the spliced node embedding vector and the feature data into a risk assessment module of the cognitive impairment risk assessment model, and obtaining a cognitive impairment risk assessment result of the target patient through processing by the risk assessment module.

3. The method of claim 2, wherein, The method comprises: inputting all the node embedding vectors into the fusion module of the cognitive impairment risk assessment model; performing average aggregation processing or splicing processing on all the node embedding vectors by using the fusion module to obtain a spliced node embedding vector.

4. The method of claim 2, wherein, The method comprises: inputting the spliced node embedding vector and the feature data into the risk assessment module of the cognitive impairment risk assessment model, and performing splicing processing on the spliced node embedding vector and the feature data by using a linear transformation layer in the risk assessment module to obtain a fusion embedding vector; inputting the fusion embedding vector into a classifier of the risk assessment module, and obtaining classification information of the target patient through processing by the classifier; obtaining a cognitive impairment risk assessment result of the target patient based on the classification information of the target patient. The classifier is a multilayer perceptron; the method comprises:

5. The method of claim 4, wherein, inputting the fusion embedding vector into an input layer of the multilayer perceptron, and transmitting the fusion embedding vector to an intermediate hidden layer of the multilayer perceptron through the input layer; performing non-linear transformation on the fusion embedding vector by using each neuron in the intermediate hidden layer to obtain intermediate hidden layer output data, and transmitting the intermediate hidden layer output data to an output layer of the multilayer perceptron; and performing non-linear transformation on the fusion embedding vector by using each neuron in the intermediate hidden layer to obtain intermediate hidden layer output data, and transmitting the intermediate hidden layer output data to an output layer of the multilayer perceptron. The output layer is used to process the intermediate hidden layer output data to obtain classification information of the target patient, the classification information being a cognitive impairment category and a probability distribution thereof.

6. The method of claim 5, wherein, The cognitive impairment risk assessment result of the target patient is obtained based on the classification information of the target patient, including: Screening the cognitive impairment category with the highest probability in the classification information of the target patient; The cognitive impairment category with the highest probability screened is taken as the cognitive impairment risk assessment result of the target patient.

7. The method of claim 1, wherein, The training process of the cognitive impairment risk assessment model includes: constructing a training data set; constructing an initial cognitive impairment risk assessment model; inputting the training data set into the initial cognitive impairment risk assessment model to be trained, and obtaining an initial cognitive impairment risk assessment result through processing by the initial cognitive impairment risk assessment model; calculating a loss function based on the initial cognitive impairment risk assessment result and a preset label, minimizing the loss function through an optimization algorithm, and updating model parameters; iteratively updating the model parameters until an end condition is reached to obtain a trained cognitive impairment risk assessment model.

8. A personal cognitive impairment risk assessment device, characterized by, including: a data acquisition module configured to acquire target data, the target data being detection data, feature data, and medical record data of a target patient; a risk assessment module configured to input the target data into a trained cognitive impairment risk assessment model, and obtain a cognitive impairment risk assessment result of the target patient through processing by the cognitive impairment risk assessment model.

9. A terminal, characterized by comprising: The terminal includes a memory, a processor, and a personal cognitive impairment risk assessment program stored in the memory and executable on the processor. When the processor executes the personal cognitive impairment risk assessment program, the steps of the personal cognitive impairment risk assessment method according to any one of claims 1-7 are implemented.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a personal cognitive impairment risk assessment program that can be executed to implement the steps of the personal cognitive impairment risk assessment method according to any one of claims 1-7.

Images