Training method, recognition method and system of cognitive impairment recognition model

Abstract

The invention discloses a training method, a recognition method and a system for a cognitive impairment recognition model, and the training method comprises the steps: obtaining a sleep training data set; performing importance feature screening on the sleep training data set to obtain a first intermediate feature set, and performing regression feature screening on the sleep training data set to obtain a second intermediate feature set; performing feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set; and inputting the target feature set into an initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model. According to the training method, a cognitive impairment recognition model can be provided, and the cognitive impairment recognition efficiency and accuracy can be improved. The invention relates to the technical field of machine learning.

Description

Technical Field

[0001] The present invention relates to the field of machine learning technology, and in particular to a training method, an identification method and a system for a cognitive impairment identification model. Background Art

[0002] Sleep is a vital process for maintaining normal physiological function in the human body and is closely related to brain activities such as cognitive function, memory consolidation, and emotional regulation. In real life, obstructive sleep apnea (OSA) is one of the most common sleep disorders. Its pathological characteristics include chronic intermittent hypoxia and disrupted sleep architecture, and are closely associated with cognitive impairment.

[0003] Currently, relevant technologies usually use cognitive impairment assessment tools such as the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment (MoCA), or the Addenbrooke's Cognitive Examination-III (ACE-III) to assess whether OSA patients have cognitive impairment. This method has high requirements for the cooperation of the subjects and is affected by human subjective factors. The scope of application of cognitive impairment assessment is limited, and the efficiency and accuracy of cognitive impairment identification are unsatisfactory.

[0004] Therefore, the problems existing in related technologies still need to be solved and optimized urgently. Summary of the Invention

[0005] The purpose of the present invention is to solve one of the technical problems existing in the related art to at least a certain extent.

[0006] To this end, an object of an embodiment of the present invention is to provide a training method, an identification method and a system for a cognitive impairment identification model, wherein the training method can provide a cognitive impairment identification model, which is conducive to improving the efficiency and accuracy of cognitive impairment identification.

[0007] In order to achieve the above technical objectives, the technical solutions adopted in the embodiments of the present application include:

[0008] In a first aspect, an embodiment of the present application provides a method for training a cognitive impairment recognition model, comprising:

[0009] Get the sleep training dataset;

[0010] Performing importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and performing regression feature screening on the sleep training dataset to obtain a second intermediate feature set;

[0011] Performing feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set;

[0012] The target feature set is input into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

[0013] In addition, the method according to the above embodiment of the present application may also have the following additional technical features:

[0014] Furthermore, in one embodiment of the present application, obtaining a sleep training data set includes:

[0015] Get the original sleep dataset;

[0016] performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information for each sleep event data, wherein the staging type of each sleep staging information is different from the staging type of the previous sleep staging information;

[0017] Performing structural parameter analysis on all the sleep stage information to obtain a sleep structural parameter group corresponding to each sleep event data;

[0018] The sleep training data set is obtained according to all the sleep structure parameter groups.

[0019] Furthermore, in one embodiment of the present application, performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information of each sleep event data includes:

[0020] Obtaining the start time information, first stage status information and first stage time information of the current sleep event data;

[0021] According to the start time information, performing sliding window stage processing on the current sleep event data to obtain second stage state information and second stage time information;

[0022] Comparing the first installment status information with the second installment status information to obtain an installment status verification result;

[0023] If the result of the stage status verification is that the first stage status information and the second stage status information are the same, the first stage time information and the start time information are updated according to the second stage time information, and the first stage status information is updated according to the second stage status information, and then the process returns to execute the step of obtaining the start time information, first stage status information and first stage time information of the current sleep event data; or, if the result of the stage status verification is that the first stage status information and the second stage status information are different, the sleep stage information is obtained according to the first stage status information and the first stage time information.

[0024] Furthermore, in one embodiment of the present application, the importance feature screening of the sleep training dataset to obtain a first intermediate feature set includes:

[0025] Constructing a random forest model and obtaining a shadow dataset corresponding to the sleep training dataset;

[0026] Inputting the shadow dataset and the sleep training dataset into the random forest model to perform importance feature classification, thereby obtaining several significant features output by the random forest model;

[0027] The first intermediate feature set is obtained according to all the significant features.

[0028] Furthermore, in one embodiment of the present application, the sleep training dataset is subjected to regression feature screening to obtain a second intermediate feature set, including:

[0029] Build regression models;

[0030] Inputting the sleep training data set into the regression model for regression training to obtain the optimal regularization parameter of the regression model;

[0031] According to the optimal regularization parameter and the regression model, non-zero coefficient features are screened for the sleep training dataset to obtain the second intermediate feature set.

[0032] Furthermore, in one embodiment of the present application, inputting the target feature set into an initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model includes:

[0033] Inputting the target feature set into a plurality of different initialized cognitive impairment recognition models for training to obtain a plurality of trained intermediate recognition models;

[0034] Performing model performance analysis on all the trained intermediate recognition models to obtain performance analysis information of each of the trained intermediate recognition models;

[0035] Based on all the performance analysis information, model screening is performed on all the trained intermediate recognition models to obtain the trained cognitive impairment recognition model.

[0036] In a second aspect, an embodiment of the present application provides a method for identifying a cognitive impairment recognition model, comprising:

[0037] Get the user's sleep detection data;

[0038] The sleep detection data is input into the above-mentioned trained cognitive impairment recognition model for recognition, to obtain the cognitive impairment recognition result of the user.

[0039] In a third aspect, an embodiment of the present application provides a training system for a cognitive impairment recognition model, comprising:

[0040] A first processing unit is used to obtain a sleep training data set;

[0041] a second processing unit, configured to perform importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and perform regression feature screening on the sleep training dataset to obtain a second intermediate feature set;

[0042] a third processing unit, configured to perform feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set;

[0043] The fourth processing unit is configured to input the target feature set into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

[0044] In a fourth aspect, an embodiment of the present application further provides an electronic device, including:

[0045] at least one processor;

[0046] at least one memory for storing at least one program;

[0047] When the at least one program is executed by the at least one processor, the at least one processor implements the above method.

[0048] In a fifth aspect, an embodiment of the present application further provides a computer-readable storage medium, which stores a program executable by a processor, and the program executable by the processor is used to implement the above method when executed by the processor.

[0049] The advantages and benefits of this application will be partially given in the following description, and partially become apparent from the following description, or learned through practice of this application:

[0050] The embodiments of the present application disclose a training method, an identification method, and a system for a cognitive impairment recognition model, wherein the training method obtains a sleep training data set; performs importance feature screening on the sleep training data set to obtain a first intermediate feature set, and performs regression feature screening on the sleep training data set to obtain a second intermediate feature set; performs feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set; and inputs the target feature set into an initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model. The training method performs importance feature screening and regression feature screening on the sleep training data set, and extracts the target feature set of the first intermediate feature set and the second intermediate feature set based on the intersection. It can fully explore the relationship between sleep training data and cognitive impairment, thereby effectively improving the accuracy of cognitive impairment recognition of the cognitive impairment recognition model obtained by subsequent training, which is beneficial to improving the efficiency of cognitive impairment recognition. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following introduction is made to the drawings of the embodiments of the present application or the related technical solutions in the prior art. It should be understood that the drawings introduced below are only for the convenience of clearly expressing some embodiments of the technical solutions of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without any creative work.

[0052] Figure 1 A flowchart of a method for training a cognitive impairment recognition model provided in an embodiment of the present application;

[0053] Figure 2 A schematic diagram of the structural framework of a training system for a cognitive impairment recognition model provided in an embodiment of the present application;

[0054] Figure 3 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0055] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present application and are not to be construed as limiting the present application. For the step numbers in the following embodiments, they are provided only for the convenience of explanation and are not intended to limit the order of the steps. The order of execution of the steps in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0057] Currently, relevant technologies usually use cognitive impairment assessment tools such as the Mini-Mental State Examination (MMSE), the Montreal Cognitive Assessment (MoCA), or the Addenbrooke's Cognitive Examination-III (ACE-III) to assess whether OSA patients have cognitive impairment. This method has high requirements for the cooperation of the subjects and is affected by human subjective factors. The scope of application of cognitive impairment assessment is limited, and the efficiency and accuracy of cognitive impairment identification are unsatisfactory.

[0058] There are also some related technologies based on polysomnography (PSG) technology. These technologies monitor multiple physiological indicators of the subject throughout the night, including electroencephalogram (EEG), oculoculogram (Oculo-oculogram), electromyogram (EMG), electrocardiogram (ECG), nasal pressure and airflow, pulse oxygen saturation (O2), and chest and abdominal respiratory movements. These indicators are then directly input into a neural network model for cognitive impairment identification. This approach typically involves simple preprocessing of multiple physiological indicators, such as removing outliers, and fails to fully utilize the complex relationship between changes in sleep event structure information and cognitive impairment, resulting in unsatisfactory accuracy in cognitive impairment identification.

[0059] It should be noted that the above-mentioned related technologies are only used to assist in understanding the technical solutions of this application and do not mean that they belong to the disclosed prior art.

[0060] In view of this, an embodiment of the present invention provides a training method, an identification method and a system for a cognitive impairment recognition model, wherein the training method performs sliding window stage processing on each sleep event data, which can fully capture the structural change information of the sleep event and record the captured sleep event structural change information in the sleep stage information; then, the method performs importance feature screening and regression feature screening on the sleep training data set respectively, and extracts the target feature set of the first intermediate feature set and the second intermediate feature set based on the intersection, which can fully explore the relationship between the sleep training data and cognitive impairment, thereby effectively improving the cognitive impairment recognition accuracy of the cognitive impairment recognition model obtained by subsequent training, which is conducive to improving the efficiency of cognitive impairment recognition.

[0061] Reference Figure 1 In an embodiment of the present application, a method for training a cognitive impairment recognition model includes:

[0062] Step 110: Obtain a sleep training data set;

[0063] In an embodiment of the present application, the structural parameters of each sleep event data can be extracted based on a number of sleep event data publicly available on the Internet; or, with the subject's authorization, the structural parameters of multiple sleep event data obtained from historical tests of the subject can be determined as a sleep training data set.

[0064] In some embodiments, obtaining a sleep training dataset includes:

[0065] Get the original sleep dataset;

[0066] performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information for each sleep event data, wherein the staging type of each sleep staging information is different from the staging type of the previous sleep staging information;

[0067] Furthermore, performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information of each sleep event data includes:

[0068] Obtaining the start time information, first stage status information and first stage time information of the current sleep event data;

[0069] According to the start time information, performing sliding window stage processing on the current sleep event data to obtain second stage state information and second stage time information;

[0070] Comparing the first installment status information with the second installment status information to obtain an installment status verification result;

[0071] If the result of the stage status verification is that the first stage status information and the second stage status information are the same, the first stage time information and the start time information are updated according to the second stage time information, and the first stage status information is updated according to the second stage status information, and then the process returns to execute the step of obtaining the start time information, first stage status information and first stage time information of the current sleep event data; or, if the result of the stage status verification is that the first stage status information and the second stage status information are different, the sleep stage information is obtained according to the first stage status information and the first stage time information.

[0072] In an embodiment of the present application, the original sleep data set includes several sleep event data, which can be data publicly available on the Internet or data obtained by the subject in historical tests. The sleep event data can specifically include apnea-hypopnea index (AHI), mean blood oxygen saturation (Mean SpO2), minimum blood oxygen saturation (LSa02), proportion of time with blood oxygen saturation ≤ 90% (CT90), total sleep time (TST), sleep efficiency (SE), proportion of time in each sleep period, latency of each sleep period and other structural parameters. The embodiment of the present application does not elaborate on the meaning of the specific structural parameters.

[0073] It can be understood that the start time information is used to indicate the start time (i.e., the first frame timestamp) of the sliding window processing the current sleep time data; the stage status information is used to indicate the sleep stage type corresponding to the time period within the sliding window. There can be multiple specific types of sleep stage types. For example, it can be divided into non-rapid eye movement sleep (NREM) and rapid eye movement sleep (REM), and NREM can be divided into falling asleep (N1), light sleep (N2), moderate sleep (N3), and deep sleep (N4) based on the four-stage classification method; or, NREM can be divided into falling asleep (N1), light sleep (N2), and deep sleep stages based on the three-stage combined classification method. The examples in this application are for illustration only and do not limit this application. The stage time information is used to indicate the start time and end time (i.e., the last frame timestamp) of the corresponding sleep stage type.

[0074] Specifically, for a sleep event data in the original sleep data set, several sleep stage information of the sleep event data can be obtained in a cyclic manner. If the current number of cycles is 1, the start time information of the sleep event data can be the starting position of the effective sleep of the sleep event data, and its first stage status information can be the same as the second stage status information under the current number of cycles; or, if the current number of cycles is greater than or equal to 2, the start time information of the sleep event data can be the end time of the second stage time information of the previous number of cycles, and the first stage time information can be the earliest start time and latest end time of several consecutive sleep stage types of the same type.

[0075] For example, in a specific loop, after obtaining the first stage status information and the second stage status information for the current sleep event data, the sleep stage types of the first stage status information and the second stage status information can be compared to determine if they are identical, thereby obtaining a stage status verification result. If the sleep stage type indicated by the first stage status information is identical to the sleep stage type indicated by the second stage status information, then the time period for the current loop and the time period for the previous loop belong to the same consecutive sleep stage. In this case, a stage status verification result can be generated indicating that the first stage status information and the second stage status information are identical. The start time information for the current loop is then updated to the end time in the second stage time information, thereby completing the update of the start time information and causing the sliding window to process the sleep time data frame by frame. Simultaneously, the end time in the first stage status information is replaced with the end time in the second stage time information, thereby completing the update of the first stage time information, so that the first stage time information indicates the earliest start time and latest end time of consecutive identical sleep stage types. The process then returns to obtaining the start time information, first stage status information, and first stage time information for the current sleep event data, and executing the next loop.

[0076] It is worth mentioning that if the sleep stage type indicated by the first stage status information is different from the sleep stage type indicated by the second stage status information, it means that the time period under the current cycle number and the time period under the previous cycle number do not belong to the same continuous sleep stage. At this time, a stage status verification result can be generated to indicate that the first stage status information and the second stage status information are different; then, the sleep stage information is directly constructed based on the obtained first stage status information and the first stage time information.

[0077] It should be noted that while constructing the sleep stage information, the second stage status information can also be determined as the first stage status information of the next cycle process, and the second stage time information can be determined as the first stage time information of the next cycle process. The remaining cycle processes are similar, and this application will not go into details here.

[0078] Performing structural parameter analysis on all the sleep stage information to obtain a sleep structural parameter group corresponding to each sleep event data;

[0079] The sleep training data set is obtained according to all the sleep structure parameter groups.

[0080] In an embodiment of the present application, after obtaining a plurality of sleep stage information of each sleep event data, structural parameter analysis can be performed on all sleep stage information of each sleep event data respectively to obtain a plurality of structural parameter groups, each structural parameter group corresponding to a sleep event data. Specifically, since the sleep stage information can indicate a sleep segment of a sleep stage type, the structural parameter group of the sleep event data may include at least one of the structural parameters such as the number of sleep segments, the duration of a single sleep segment, the interval segments between the same sleep segment, and the interval duration between the same sleep segments. Among them, the number of sleep segments can be the number of occurrences of sleep segments of each sleep stage type of the sleep event data; the single sleep segment duration can be the average duration of the sleep segments of each sleep stage type of the sleep event data; the interval segments between the same sleep segment can be the number of sleep segments of other sleep stage types that appear between two sleep segments of the same sleep stage type; the interval duration between the same sleep segment can be the duration of sleep segments of other sleep stage types that appear between two sleep segments of the same sleep stage type. The sleep structural parameter groups of the remaining sleep event data are similar and can be simply deduced by analogy.

[0081] It is understandable that after obtaining the sleep structure parameter groups of all sleep event data, a corresponding multidimensional feature matrix can be constructed based on all sleep structure parameter groups, and the multidimensional feature matrix can be determined as a sleep training dataset.

[0082] Step 120: performing importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and performing regression feature screening on the sleep training dataset to obtain a second intermediate feature set;

[0083] In an embodiment of the present application, the sleep training dataset can be screened based on the importance of the features to obtain a first intermediate feature set; and several regression features in the sleep training dataset can be extracted based on the Lasso regression algorithm to obtain a second intermediate feature set.

[0084] In some embodiments, the performing importance feature screening on the sleep training dataset to obtain a first intermediate feature set includes:

[0085] Constructing a random forest model and obtaining a shadow dataset corresponding to the sleep training dataset;

[0086] Inputting the shadow dataset and the sleep training dataset into the random forest model to perform importance feature classification, thereby obtaining several significant features output by the random forest model;

[0087] The first intermediate feature set is obtained according to all the significant features.

[0088] In an embodiment of the present application, since the sleep training data set includes several structural parameter groups, each structural parameter group includes several structural parameters, each structural parameter can be used as a training feature in the sleep training data set; then a random permutation (Shuffle) operation is performed on each training feature to generate a shadow feature corresponding to each training feature, and a shadow data set is determined based on all the shadow features.

[0089] It can be understood that the importance feature classification can be based on the Boruta algorithm, which inputs the shadow dataset and the sleep training dataset into the random forest model for importance prediction. Through multiple rounds of iterations, several training features with importance scores greater than the importance scores of the shadow features are obtained, and the training features with importance scores greater than the importance scores of the shadow features are recorded as significant features. Then, based on all the significant features obtained, the first intermediate feature set is constructed.

[0090] In some embodiments, performing regression feature screening on the sleep training dataset to obtain a second intermediate feature set includes:

[0091] Build regression models;

[0092] Inputting the sleep training data set into the regression model for regression training to obtain the optimal regularization parameter of the regression model;

[0093] According to the optimal regularization parameter and the regression model, non-zero coefficient features are screened for the sleep training dataset to obtain the second intermediate feature set.

[0094] In an embodiment of the present application, a regression model can be constructed based on the Lasso regression algorithm, and the sleep training data set is input into the regression model for training, and the optimal regularization parameter of the regression model is determined by the K-fold cross-validation method; then, based on the optimal regularization parameter, the regression model can output several non-zero coefficients, each non-zero coefficient corresponds to a training feature selected by the regression model in the sleep training data set, and a second intermediate feature set can be constructed based on the training features corresponding to each non-zero coefficient.

[0095] Step 130: Perform feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set;

[0096] In an embodiment of the present application, feature intersection extraction can be to perform an intersection operation on the first intermediate feature set and the second intermediate feature set to obtain a number of training features that coexist in the first intermediate feature set and the second intermediate feature set, and determine the coexisting several training features as the target feature set.

[0097] Step 140: Input the target feature set into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

[0098] In an embodiment of the present application, the target feature set can be input into an initialized cognitive loss recognition model for model training. The cognitive loss recognition model can be any one of the machine learning models such as support vector machine (SVM), gradient boosting model (such as XGBoost, GBM), logistic regression model, multi-layer perceptron (MLP), etc., so as to obtain a trained cognitive impairment recognition model.

[0099] In some embodiments, inputting the target feature set into an initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model includes:

[0100] Inputting the target feature set into a plurality of different initialized cognitive impairment recognition models for training to obtain a plurality of trained intermediate recognition models;

[0101] Performing model performance analysis on all the trained intermediate recognition models to obtain performance analysis information of each of the trained intermediate recognition models;

[0102] Based on all the performance analysis information, model screening is performed on all the trained intermediate recognition models to obtain the trained cognitive impairment recognition model.

[0103] In an embodiment of the present application, the target feature sets can be input into different initialized cognitive impairment recognition models for training. For example, the target feature sets can be input into support vector machines, gradient boosting models, and multi-layer perceptrons for training to obtain trained intermediate recognition models. The performance of each trained intermediate recognition model is then analyzed to obtain performance analysis information of each trained intermediate recognition model. The performance analysis information can specifically be performance indicators such as ROC curves, AUC curves, and accuracy rates.

[0104] It is understandable that after obtaining the performance analysis information of each trained intermediate recognition model, the trained intermediate recognition model corresponding to the optimal performance analysis information can be determined as the trained cognitive impairment recognition model. In addition, for machine learning models, the accuracy of the model prediction results can be measured by a loss function. The loss function is defined on a single training data and is used to measure the prediction error of a training data. Specifically, the loss value of the training data is determined by the label of the single training data and the prediction result of the model for the training data. In actual training, a training data set has a lot of training data, so a cost function is generally used to measure the overall error of the training data set. The cost function is defined on the entire training data set and is used to calculate the average value of the prediction error of all training data, which can better measure the prediction effect of the model. For general machine learning models, based on the aforementioned cost function, plus a regularization term that measures the complexity of the model, it can be used as the objective function of training. Based on this objective function, the loss value of the entire training data set can be calculated. There are many types of commonly used loss functions, such as 0-1 loss function, square loss function, absolute loss function, logarithmic loss function, cross entropy loss function, etc., which can all be used as loss functions of machine learning models, which will not be elaborated here one by one. In the embodiment of the present application, any one of the loss functions can be selected to determine the loss value of the training, such as the cross entropy loss function. Based on the loss value of the training, the back propagation algorithm is used to update the parameters of the model, and a trained intermediate recognition model can be obtained by iterating several rounds. The specific number of iterations can be pre-set, or the training is considered to be completed when the test set meets the accuracy requirements.

[0105] In an embodiment of the present application, a method for identifying a cognitive impairment identification model includes:

[0106] Obtain the user's sleep detection data;

[0107] The sleep detection data is input into the trained cognitive impairment recognition model for recognition, to obtain a cognitive impairment recognition result of the user.

[0108] In an embodiment of the present application, the user's sleep detection data can be data directly collected based on polysomnography (PSG) technology; or, it can also be obtained by performing events staging, structural parameter analysis and other operations on the collected data based on the data collected based on PSG technology.

[0109] It is understandable that in actual applications, the user's sleep detection data can be input into a trained cognitive impairment recognition model, and the sleep detection data can be identified by the trained cognitive impairment recognition model to obtain the user's cognitive impairment recognition result. The cognitive impairment recognition result is used to characterize the user's possible cognitive impairment and the probability value of the cognitive impairment.

[0110] It should be noted that, according to experimental data, when a multi-layer perceptron is used for the cognitive impairment recognition model, the features obtained by the Lasso regression algorithm are directly used to train the model. The accuracy of the trained model is 0.802, and the AUC index value is 0.84. The embodiment of the present application is based on the importance and regression feature extraction of the sleep training data machine containing the sleep structure parameter group, and the target feature set constructed by the intersection is used to train the model. The accuracy of the trained model is 0.825 and the AUC is 0.89. It can better screen parameter factors related to cognitive impairment, reduce the interference of other parameters bianl2, and make the performance of the cognitive impairment recognition model better.

[0111] A training system for a cognitive impairment recognition model proposed in accordance with an embodiment of the present application is described in detail below with reference to the accompanying drawings.

[0112] Reference Figure 2 A training system for a cognitive impairment recognition model proposed in an embodiment of the present application includes:

[0113] A first processing unit 101 is configured to obtain a sleep training data set;

[0114] a second processing unit 102, configured to perform importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and perform regression feature screening on the sleep training dataset to obtain a second intermediate feature set;

[0115] The third processing unit 103 is configured to perform feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set;

[0116] The fourth processing unit 104 is configured to input the target feature set into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

[0117] It can be understood that the contents of the above method embodiments are all applicable to the present system embodiments, the functions specifically implemented by the present system embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0118] Reference Figure 3 , an embodiment of the present application further provides an electronic device, including:

[0119] at least one processor 201;

[0120] At least one memory 202, configured to store at least one program;

[0121] When the at least one program is executed by the at least one processor 201 , the at least one processor 201 implements the above method embodiment.

[0122] Similarly, it can be understood that the contents of the above method embodiments are applicable to the present device embodiments, the functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0123] An embodiment of the present application further provides a computer-readable storage medium, in which a program executable by the processor 201 is stored. The program executable by the processor 201 is used to implement the above-mentioned method embodiment when executed by the processor 201.

[0124] Similarly, the contents of the above method embodiments are applicable to the computer-readable storage medium embodiments. The functions specifically implemented by the computer-readable storage medium embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0125] In some optional embodiments, the functions / operations mentioned in the block diagram may not occur in the order mentioned in the operation diagram. For example, depending on the functions / operations involved, the two boxes shown in succession may actually be executed substantially simultaneously or the boxes can sometimes be executed in reverse order. In addition, the embodiments presented and described in the flow chart of the present application are provided in an exemplary manner for the purpose of providing a more comprehensive understanding of the technology. The disclosed method is not limited to the operations and logic flows presented herein. Optional embodiments are contemplated in which the order of the various operations is changed and the sub-operations described as a part of a larger operation are performed independently.

[0126] In addition, although the present application is described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in separate physical devices or software modules. It is also understood that a detailed discussion of the actual implementation of each module is not necessary for understanding the present application. More specifically, given the properties, functions, and internal relationships of the various functional modules in the devices disclosed herein, the actual implementation of the module will be understood within the routine skills of an engineer. Therefore, a person skilled in the art can implement the present application as set forth in the claims using ordinary techniques without undue experimentation. It is also understood that the specific concepts disclosed are merely illustrative and are not intended to limit the scope of the present application, which is determined by the full scope of the appended claims and their equivalents.

[0127] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the embodiment method of the present application. The aforementioned storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0128] The logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as an ordered list of executable instructions for implementing the logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (e.g., a computer-based system, a system including a processor, or other system that can fetch and execute instructions from an instruction execution system, apparatus, or device). For purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0129] More specific examples (a non-exhaustive list) of computer-readable media include the following: an electrical connection with one or more wires (electronic devices), a portable computer disk cartridge (magnetic devices), a random access memory (RAM), a read-only memory (ROM), an erasable and programmable read-only memory (EPROM or flash memory), a fiber optic device, and a portable compact disc read-only memory (CDROM). In addition, the computer-readable medium may even be paper or other suitable medium on which the program is printed, since the program may be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, deciphering, or processing in another suitable manner as necessary, and then stored in a computer memory.

[0130] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used to implement: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array (PGA), a field programmable gate array (FPGA), etc.

[0131] In the above description of this specification, reference to the terms "one embodiment / example," "another embodiment / example," or "certain embodiments / examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in any appropriate manner in any one or more embodiments or examples.

[0132] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and intent of the present application, and that the scope of the present application is defined by the claims and their equivalents.

[0133] The above is a specific description of the preferred implementation of the present application, but the present application is not limited to the embodiments. Those skilled in the art may make various equivalent modifications or substitutions without violating the spirit of the present application, and these equivalent modifications or substitutions are all included in the scope defined by the claims of the present application.

Claims

1. A method for training a cognitive impairment recognition model, characterized in that: include: Get the sleep training dataset; Performing importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and performing regression feature screening on the sleep training dataset to obtain a second intermediate feature set; Performing feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set; The target feature set is input into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

2. The method according to claim 1, characterized in that The step of obtaining a sleep training data set includes: Get the original sleep dataset; Performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information for each sleep event data, wherein the staging type of each sleep staging information is different from the staging type of the previous sleep staging information; Performing structural parameter analysis on all the sleep stage information to obtain a sleep structural parameter group corresponding to each sleep event data; The sleep training data set is obtained according to all the sleep structure parameter groups.

3. The method according to claim 2, characterized in that The step of performing event staging on each sleep event data in the original sleep data set to obtain a plurality of sleep staging information of each sleep event data includes: Obtaining the start time information, first stage status information and first stage time information of the current sleep event data; According to the start time information, performing sliding window stage processing on the current sleep event data to obtain second stage state information and second stage time information; Comparing the first installment status information with the second installment status information to obtain an installment status verification result; If the result of the stage status verification is that the first stage status information and the second stage status information are the same, the first stage time information and the start time information are updated according to the second stage time information, and the first stage status information is updated according to the second stage status information, and then the process returns to execute the step of obtaining the start time information, first stage status information and first stage time information of the current sleep event data; or, if the result of the stage status verification is that the first stage status information and the second stage status information are different, the sleep stage information is obtained according to the first stage status information and the first stage time information.

4. The method according to claim 1, wherein The step of screening the sleep training dataset for important features to obtain a first intermediate feature set includes: Constructing a random forest model and obtaining a shadow dataset corresponding to the sleep training dataset; Inputting the shadow dataset and the sleep training dataset into the random forest model to perform importance feature classification, thereby obtaining several significant features output by the random forest model; The first intermediate feature set is obtained according to all the significant features.

5. The method according to claim 1, wherein The step of performing regression feature screening on the sleep training dataset to obtain a second intermediate feature set includes: Build regression models; Inputting the sleep training data set into the regression model for regression training to obtain the optimal regularization parameter of the regression model; According to the optimal regularization parameter and the regression model, non-zero coefficient features are screened for the sleep training dataset to obtain the second intermediate feature set.

6. The method according to claim 1, characterized in that Inputting the target feature set into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model includes: Inputting the target feature set into a plurality of different initialized cognitive impairment recognition models for training to obtain a plurality of trained intermediate recognition models; Performing model performance analysis on all the trained intermediate recognition models to obtain performance analysis information of each of the trained intermediate recognition models; Based on all the performance analysis information, model screening is performed on all the trained intermediate recognition models to obtain the trained cognitive impairment recognition model.

7. A method for identifying a cognitive impairment identification model, characterized in that: include: Get the user's sleep detection data; The sleep detection data is input into the trained cognitive impairment recognition model according to any one of claims 1 to 6 for recognition, to obtain a cognitive impairment recognition result of the user.

8. A training system for a cognitive impairment recognition model, characterized in that: include: A first processing unit is used to obtain a sleep training data set; a second processing unit, configured to perform importance feature screening on the sleep training dataset to obtain a first intermediate feature set, and perform regression feature screening on the sleep training dataset to obtain a second intermediate feature set; a third processing unit, configured to perform feature intersection extraction on the first intermediate feature set and the second intermediate feature set to obtain a target feature set; The fourth processing unit is configured to input the target feature set into the initialized cognitive impairment recognition model for model training to obtain a trained cognitive impairment recognition model.

9. An electronic device, characterized in that: include: at least one processor; at least one memory for storing at least one program; When the at least one program is executed by the at least one processor, the at least one processor implements the method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a program executable by a processor, characterized in that: The program executable by the processor is used to implement the method according to any one of claims 1 to 7 when executed by the processor.

Images