Mild cognitive impairment patient screening method and device based on machine learning model, storage medium and electronic equipment

Abstract

The invention provides a mild cognitive impairment patient screening method and device based on a machine learning model, a storage medium and electronic equipment, and relates to the technical field of mild cognitive impairment screening. On the basis of a literature review method, a professional domain knowledge injection method, a domain expert consultation method and a Shap interpretable machine learning method, screening out characteristic variables related to mild cognitive impairment from the evaluation data to serve as training data; taking the optimal AUC as a training direction, training an extreme gradient lifting algorithm model based on the training data, and obtaining an alternative model; constructing an external independent verification data set; verifying the alternative model based on the external independent verification data set; and selecting the alternative model with the best verification effect as a rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment. According to the technical scheme provided by the invention, the mild cognitive impairment can be quickly and accurately screened.

Description

Technical Field

[0001] The present application relates to the technical field of mild cognitive impairment screening, and specifically to a method, device, storage medium and electronic device for screening patients with mild cognitive impairment based on a machine learning model. Background Art

[0002] Currently, medical resources in my country are relatively scarce. Community screening for MCI (mild cognitive impairment) requires face-to-face assessments by professionals. Although electronic questionnaire assessment systems have been implemented, they are still essentially questionnaire-based. Whether using paper or electronic questionnaires, these assessments present significant challenges, including the necessary human, material, and financial resources, as well as professional and technical barriers (evaluators must undergo repeated training and assessment). This, to a certain extent, limits the key point of early screening for cognitive impairment and is a significant constraint on the implementation of universal cognitive function screening among middle-aged and elderly community residents.

[0003] With the rapid development of artificial intelligence, the use of machine learning methods for cognitive impairment prediction research is becoming increasingly popular. However, existing technologies face challenges such as the difficulty and high cost of collecting patient characteristics, generally suboptimal model efficiency for MCI (most with an AUC below 0.75), and a lack of regional external validation. Summary of the Invention

[0004] The embodiments of the present application provide a method, device, storage medium and electronic device for screening patients with mild cognitive impairment based on a machine learning model to solve the technical problems existing in the prior art.

[0005] Other features and advantages of the present application will become apparent from the following detailed description, or may be learned in part by practice of the present application.

[0006] According to a first aspect of an embodiment of the present application, a method for screening patients with mild cognitive impairment based on a machine learning model is provided, comprising: Construct a cognitive health cohort of natural populations in the community and obtain cognitive-related assessment data of cohort study samples; Based on literature review, professional domain knowledge injection, domain expert consultation, and Shap interpretable machine learning, characteristic variables related to mild cognitive impairment are screened from the evaluation data as training data; Taking the best AUC as the training direction, training the extreme gradient boosting algorithm model based on the training data to obtain an alternative model; Construct an external independent validation dataset for external validation of the model; Validating the candidate model based on the external independent validation dataset; The alternative model with the best validation effect was selected as the rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

[0007] In some embodiments of the present application, based on the aforementioned scheme, the steps of constructing a cognitive health cohort of a natural population in a community and obtaining cognitive-related assessment data of cohort study samples include: Population samples were recruited from natural population communities. Using a cohort research strategy as the starting point, comprehensive cognitive health assessment / diabetes comprehensive assessment was conducted on all subjects. Basic demographic information, cognitive and influencing factor assessment data, physical examination and biological specimen testing information of the sample population were systematically collected.

[0008] In some embodiments of the present application, based on the aforementioned scheme, in the process of constructing a cognitive health cohort of a natural population in the community, a three-level quality control strategy is adopted to carry out quality control work at the three levels of preparation, execution, and maintenance.

[0009] In some embodiments of the present application, based on the aforementioned scheme, the method of screening characteristic variables related to mild cognitive impairment from the evaluation data based on the literature review method, the method of injecting professional domain knowledge, the method of consulting domain experts, and the Shap interpretable machine learning method as training data includes: Based on the literature review method, we collected the influencing factors related to mild cognitive impairment from the literature related to mild cognitive impairment and formed a review list of influencing factors of mild cognitive impairment; Screening characteristic variables closely related to the occurrence of cognitive impairment from the evaluation data based on the review list of factors affecting mild cognitive impairment; An evaluation model is developed based on the review list of influencing factors of mild cognitive impairment, wherein the evaluation indicators of the evaluation model include: blood collection indicators, lifestyle characteristics, scale assessment characteristics, cognitive assessment indicators, previous chronic disease characteristics, and physical dimension indicators; Conducting pre-experimental verification of the evaluation model based on expert experience and evaluation scenarios; The validated assessment model was used to evaluate the characteristic variables closely related to the occurrence of cognitive impairment, and the assessment results were cleaned and sorted to obtain a set of candidate characteristic variables related to mild cognitive impairment. Perform category classification processing on the candidate feature variable set to obtain multiple feature subsets; The features in each feature subset are screened using domain expert consultation method; The filtered feature subsets are sequentially combined into single-mode and multi-mode combinations to obtain the total candidate feature variable set; Using the Shap interpretable machine learning method, with mild cognitive impairment as the label, the features in each feature subset in the total candidate feature variable set are ranked by their importance in terms of their contribution to label prediction, and the feature distribution results of multiple groups of feature importance rankings are obtained; The feature distribution results of each group of feature importance rankings are sorted out, and combined with expert experience, the top N feature variables in each feature subset are retained as training data, where N is an integer greater than or equal to 1.

[0010] In some embodiments of the present application, based on the aforementioned scheme, the optimal AUC is used as the training direction, and an extreme gradient boosting algorithm model is trained based on the training data to obtain an alternative model, including: The optimal AUC is used as the training direction for model training based on each feature combination variable. The extreme gradient boosting algorithm model is trained based on each feature combination variable to obtain multiple preliminary models. Establish model screening criteria based on model performance; Based on the model screening conditions, a model that meets the requirements is selected from multiple preliminary models as an alternative model.

[0011] In some embodiments of the present application, based on the aforementioned scheme, constructing an external independent validation dataset for external validation of the model includes: Cognitive-related assessment data of apparently healthy middle-aged and elderly people were collected from communities that were not the source of model training data to form an external independent validation dataset.

[0012] In some embodiments of the present application, based on the aforementioned solution, the validating the candidate model based on the external independent validation dataset includes: Clean and organize the external independent validation data set, and extract the cognitive screening assessment result label data required for external validation and the alternative model feature data entering the external validation phase; The alternative models are verified one by one based on the cognitive screening assessment result label data and the alternative model feature data.

[0013] According to a second aspect of an embodiment of the present application, a device for screening patients with mild cognitive impairment based on a machine learning model is provided, comprising: The first construction unit is used to construct a cognitive health cohort of natural populations in the community and obtain cognitive-related evaluation data of cohort study samples; A first screening unit is configured to screen characteristic variables related to mild cognitive impairment from the evaluation data as training data based on a literature review method, a professional domain knowledge injection method, a domain expert consultation method, and a Shap interpretable machine learning method; A training unit, configured to train an extreme gradient boosting algorithm model based on the training data with the optimal AUC as a training direction to obtain an alternative model; The second construction unit is used to construct an external independent validation dataset for external validation of the model; a verification unit, configured to verify the candidate model based on the external independent verification data set; The second screening unit is used to select the alternative model with the best verification effect as a rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

[0014] According to a third aspect of an embodiment of the present application, a computer-readable storage medium is provided, wherein the storage medium stores computer instructions. When the computer instructions are executed on a computer, the computer executes the method according to the first aspect.

[0015] According to a fourth aspect of the embodiments of the present application, there is provided an electronic device, including: a memory and a processor; The memory is used to store computer instructions; The processor is configured to call the computer instructions stored in the memory so that the electronic device executes the method according to the first aspect.

[0016] Compared with traditional paper and electronic screening tool evaluation methods such as the MoCA scale, the technical solution of this application is faster, has more stable evaluation results (in contrast, human evaluation is more subjective), takes less time, extracts fewer features, and has more diverse application scenarios. At the same time, it can, to a certain extent, break through the dependence on human resources in scale evaluation and achieve large-scale screening. The model training and external independent verification data in the technical solution of this application are all derived from real-world natural population communities. Compared with the training sets of existing methods that mostly come from hospital patients, the application value of the technical solution of this application is higher. The model in the technical solution of this application is externally independently verified and has better generalization capabilities.

[0017] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The accompanying drawings are incorporated into and constitute a part of the specification, illustrating embodiments consistent with the present application and, together with the specification, explaining the principles of the present application. Obviously, the drawings described below are only some embodiments of the present application, and those skilled in the art can derive other drawings based on these drawings without inventive effort. In the drawings: Figure 1A schematic diagram of a process for screening patients with mild cognitive impairment based on a machine learning model according to one embodiment of the present application is shown; Figure 2 A block diagram of a device for screening patients with mild cognitive impairment based on a machine learning model according to one embodiment of the present application is shown; Figure 3 A block diagram of an electronic device according to an embodiment of the present application is shown; Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown. DETAILED DESCRIPTION

[0019] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided so that this application will be thorough and complete and will fully convey the concepts of the example embodiments to those skilled in the art.

[0020] In addition, described feature, structure or characteristic can be combined in one or more embodiments in any suitable manner.In the following description, many specific details are provided so as to provide a full understanding of the embodiments of the present application. However, it will be appreciated by those skilled in the art that the technical scheme of the present application can be put into practice without one or more of the specific details, or other methods, components, devices, steps etc. can be adopted. In other cases, known methods, devices, implementations or operations are not shown or described in detail to avoid blurring the various aspects of the application.

[0021] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically separate entities. That is, these functional entities may be implemented in software, in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0022] The flowcharts shown in the accompanying drawings are for illustrative purposes only and do not necessarily include all contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may be decomposed, while others may be combined or partially combined. Therefore, the actual execution order may vary depending on the actual situation.

[0023] It should be noted that the “plurality” mentioned in this article refers to two or more.

[0024] It should be noted that the terms "first," "second," and the like in the specification and claims of this application and the accompanying drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, such that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described.

[0025] To make the objectives, technical solutions, and advantages of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0026] The following will describe some embodiments of the present application in detail with reference to the accompanying drawings. In the absence of conflict, the following embodiments and features therein may be combined with each other.

[0027] In the research on the prediction of cognitive impairment diseases, machine learning algorithms such as random forests and support vector machines can effectively improve the prediction accuracy. This paper briefly reviews the research methods of mild cognitive impairment-related models in the existing technology.

[0028] The support vector machine (SVM) is a supervised learning model primarily used for classification and regression. It utilizes the principle of structural risk minimization, constructs a maximum-margin decision boundary, and incorporates kernel function techniques to effectively address problems such as small sample sizes, high-dimensional pattern recognition, and nonlinear classification. Currently, this algorithm is being used in cognitive impairment research, primarily focusing on cognitive prediction using features from speech, imaging, and bioelectrical signals. In recent years, this algorithm has been extensively applied to distinguishing Alzheimer's disease from mild cognitive impairment using diverse categorical features. Regarding the joint prediction of cognitive impairment with multimodal data, Ruixuan Wu et al. found that the SVM classifier achieved the best classification results, achieving an accuracy of 87%, in a study using EEG, speech, and digital cognitive methods to identify mild cognitive impairment. Yuye Wang et al. found that SVMs had good predictive performance in a study of cognitive impairment and urine metabolomics, biomarkers, and general demographic characteristics. Jiali Chai et al. used handwriting and EEG signals to screen for mild cognitive impairment, and the support vector machine algorithm demonstrated good classification results in this study. In terms of distinguishing different degrees of cognitive impairment, Fangyang Jiao et al. found in a study based on PET radiomics analysis that the SVM algorithm also has a good classification effect in distinguishing Alzheimer's disease and mild cognitive impairment. At the same time, Buhari Ibrahim et al. found through a systematic review study that the multimodal support vector machine algorithm is the most commonly used machine learning method in comparing the diagnostic ability of resting-state functional magnetic resonance imaging in identifying functional connectivity abnormalities in the default mode network of AD or MCI patients with healthy controls using machine learning methods.

[0029] Random Forest is an ensemble learning method that makes predictions by constructing and combining multiple decision trees. This model has relatively good generalization and robustness, and is capable of handling high-dimensional data and a large number of features. In recent years, this method has been increasingly used in the field of cognitive impairment prediction, with a relatively large number of studies focusing on the relationship between cognitive impairment and imaging features. For example, Juan Yang et al. used a random forest approach to establish a diagnostic model for mild cognitive impairment using basic demographic characteristics, serological indicators, and imaging markers, and achieved good results compared to the gold standard. Xiaan Bi et al. used a random forest algorithm to screen and combine imaging features with genetic characteristics to construct a model for the etiology of mild cognitive impairment, achieving good predictive results. Sajjad Fouladvand et al., after analyzing data from over 530,000 samples, found that the random forest model had good predictive ability for the onset of MCI (ROC-AUC of 68.2±0.7).

[0030] The multimodal features within existing technical models include but are not limited to cerebrospinal fluid indicators (such as tau, Aβ, etc.), imaging features (such as MRI, PET, etc.), bioelectric signals (such as EEG signals, etc.), etc. The features currently widely used in multimodal models often face the characteristics of greater difficulty in acquisition, relatively high acquisition fees / indicator detection costs, expensive data acquisition equipment or high-net-value consumables, and patients facing a higher risk of invasive / traumatic infection.

[0031] At the same time, the model's efficiency for MCI is generally unsatisfactory (most AUCs are below 0.75), the model itself lacks regional external validation, and the model's generalization ability needs further demonstration.

[0032] To address the problems existing in the existing technology, the technical solution of this application uses machine learning and interpretable machine learning reverse engineering technology to build a rapid auxiliary screening model for MCI among middle-aged and elderly residents in the community using a simple and small set of simple features. Independent external validation is then conducted among the middle-aged and elderly community population, thereby demonstrating the application potential and generalization ability of this screening model method. On this basis, it will assist in the universal and early screening of cognitive function among the middle-aged and elderly population in my country.

[0033] See also Figure 1 , shows a flow chart of a method for screening patients with mild cognitive impairment based on a machine learning model according to an embodiment of the present application.

[0034] like Figure 1 As shown, a method for screening patients with mild cognitive impairment based on a machine learning model is presented, specifically including steps S100 to S600.

[0035] refer to Figure 1 , step S100, constructing a cognitive health cohort of natural population in the community and obtaining cognitive related assessment data of cohort study samples.

[0036] In some feasible embodiments, based on the above scheme, the steps of constructing a community natural population cognitive health cohort and obtaining cognition-related assessment data of cohort study samples include: Population samples were recruited from natural population communities. Using a cohort research strategy as the starting point, comprehensive cognitive health assessment / diabetes comprehensive assessment was conducted on all subjects. Basic demographic information, cognitive and influencing factor assessment data, physical examination and biological specimen testing information of the sample population were systematically collected.

[0037] It should be noted that, in this embodiment, the data obtained are all from real-world natural population communities. Compared with the traditional method of obtaining data from hospital patients, this method has more practical application value.

[0038] In some feasible embodiments, based on the aforementioned scheme, in the process of constructing a cognitive health cohort of a natural population in the community, a three-level quality control strategy is adopted to carry out quality control work at the three levels of preparation, execution, and maintenance.

[0039] It should be noted that in this embodiment, the three-level quality control strategy is adopted in the process of constructing a cognitive health cohort of a natural population in the community, which can lay a solid foundation for accurately collecting real-world data, conducting cognitive-related scientific research, and sustainable cohort development.

[0040] Continue to refer Figure 1 In step S200, based on the literature review method, the professional domain knowledge injection method, the domain expert consultation method and the Shap interpretable machine learning method, characteristic variables related to mild cognitive impairment are screened out from the evaluation data as training data.

[0041] In some feasible embodiments, based on the aforementioned solution, the method of screening characteristic variables related to mild cognitive impairment from the evaluation data based on the literature review method, the method of professional domain knowledge injection, the method of domain expert consultation, and the Shap interpretable machine learning method as training data includes: Based on the literature review method, we collected the influencing factors related to mild cognitive impairment from the literature related to mild cognitive impairment and formed a review list of influencing factors of mild cognitive impairment; Screening characteristic variables closely related to the occurrence of cognitive impairment from the evaluation data based on the review list of factors affecting mild cognitive impairment; An evaluation model is developed based on the review list of influencing factors of mild cognitive impairment, wherein the evaluation indicators of the evaluation model include: blood collection indicators, lifestyle characteristics, scale assessment characteristics, cognitive assessment indicators, previous chronic disease characteristics, and physical dimension indicators; Conducting pre-experimental verification of the evaluation model based on expert experience and evaluation scenarios; The validated assessment model was used to evaluate the characteristic variables closely related to the occurrence of cognitive impairment, and the assessment results were cleaned and sorted to obtain a set of candidate characteristic variables related to mild cognitive impairment. Perform category classification processing on the candidate feature variable set to obtain multiple feature subsets; The features in each feature subset are screened using domain expert consultation method; The filtered feature subsets are sequentially combined into single-mode and multi-mode combinations to obtain the total candidate feature variable set; Using the Shap interpretable machine learning method, with mild cognitive impairment as the label, the features in each feature subset in the total candidate feature variable set are ranked by their importance in terms of their contribution to label prediction, and the feature distribution results of multiple groups of feature importance rankings are obtained; The feature distribution results of each group of feature importance rankings are sorted out, and combined with expert experience, the top N feature variables in each feature subset are retained as training data, where N is an integer greater than or equal to 1.

[0042] Exemplarily, N takes the value of 15. In the process of combining features in order of importance, they can be combined in the form of "top 5 feature combination", "top 6 feature combination", "top 7 feature combination", "top 8 feature combination", "top 9 feature combination", "top 10 feature combination", "top 11 feature combination", "top 12 feature combination", "top 13 feature combination", "top 14 feature combination", "top 15 feature combination", etc.

[0043] It should be noted that the “first x feature combinations” described in this implementation refer to combining the first x features, where x is a positive integer and is related to the feature data of the retained feature importance.

[0044] For example, "the top 5 feature combination" means combining the first, second, third, fourth, and fifth features among the top 15 feature variables ranked by importance.

[0045] Continue to refer Figure 1 In step S300 , the extreme gradient boosting algorithm model is trained based on the training data with the best AUC as the training direction to obtain an alternative model.

[0046] It should be noted that the AUC mentioned in this application refers to the Area Under the Curve, that is, the area under the ROC curve, which is a commonly used performance evaluation indicator used to measure the overall classification ability of the model.

[0047] In some feasible embodiments, based on the above scheme, the optimal AUC is used as the training direction, and an extreme gradient boosting algorithm model is trained based on the training data to obtain an alternative model, including: The optimal AUC is used as the training direction for model training based on each feature combination variable. The extreme gradient boosting algorithm model is trained based on each feature combination variable to obtain multiple preliminary models. Establish model screening criteria based on model performance; Based on the model screening conditions, a model that meets the requirements is selected from multiple preliminary models as an alternative model.

[0048] Exemplarily, the model screening conditions are: AUC ≥ 0.8 and F1 ≥ 0.8.

[0049] Among them, F1 refers to the F1 value (F1 - Score), which is a measurement indicator for classification problems. It is the harmonic mean of precision and recall, and is used to comprehensively evaluate the performance of the model.

[0050] Continue to refer Figure 1 , step S400, constructing an external independent validation dataset for external validation of the model.

[0051] In some feasible embodiments, based on the above scheme, constructing an external independent validation dataset for external validation of the model includes: Cognitive-related assessment data of apparently healthy middle-aged and elderly people were collected from communities that were not the source of model training data to form an external independent validation dataset.

[0052] Continue to refer Figure 1 , step S500, verifying the alternative model based on the external independent verification data set.

[0053] In some feasible embodiments, based on the above solution, the validating the candidate model based on the external independent validation dataset includes: Clean and organize the external independent validation data set, and extract the cognitive screening assessment result label data required for external validation and the alternative model feature data entering the external validation phase; The alternative models are verified one by one based on the cognitive screening assessment result label data and the alternative model feature data.

[0054] It should be noted that, in this embodiment, the model is verified by external verification, which improves the generalization ability of the model.

[0055] Continue to refer Figure 1 , step S600, selecting the candidate model with the best verification effect as the rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

[0056] In summary, the method provided in the embodiments of the present application has the following characteristics: 1. Compared with traditional paper-based or electronic screening tools such as the MoCA scale, the screening method provided by this invention is faster, has more stable evaluation results (in contrast, manual evaluation is more subjective), is less time-consuming, requires fewer features to extract, and has more diverse application scenarios. Furthermore, it can, to a certain extent, overcome the human resource dependence in scale assessment and achieve widespread screening.

[0057] 2. The model training and external independent verification data in the present invention are all derived from real-world natural population communities. Compared with the existing methods whose training sets are mostly derived from hospital patients, the application value of the present invention is higher.

[0058] 3. In the selection of features for the rapid screening model for mild cognitive impairment, the present invention uses seven innovative feature engineering features, all of which are derived from physical examination data and can be measured using a conventional tape measure. Compared with traditional scale screening, this method has better extrapolation ability. It can be collected by institutions at all levels, including but not limited to health management centers, community health service centers, cognitively friendly community screening sites, and township health centers. The collection difficulty and plan are both current mature methods, and there is no need to purchase additional large equipment. The collection thresholds for the two inquiry variables are low, and no professional training is required for evaluators. The two measurement variables do not require high-net-worth equipment / consumables.

[0059] 4. The rapid screening model for mild cognitive impairment in this invention performed better than previous models in cross-validation of the model-building dataset (AUC > 0.8). Furthermore, this screening model was independently validated externally in different years and on a population different from the model-building dataset, achieving an AUC of 0.9 and an F1 of 84.86%, demonstrating the excellent generalization ability of this rapid screening model for mild cognitive impairment.

[0060] The following describes an embodiment of the device of the present application, which can be used to implement a method for screening patients with mild cognitive impairment based on a machine learning model in the above-mentioned embodiment of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0061] Reference Figure 2 As shown, according to one embodiment of the present application, a device 200 for screening patients with mild cognitive impairment based on a machine learning model includes: The first construction unit 201 is used to construct a cognitive health cohort of a natural population in the community and obtain cognitive related evaluation data of the cohort study samples; A first screening unit 202 is configured to screen characteristic variables related to mild cognitive impairment from the evaluation data as training data based on a literature review method, a professional domain knowledge injection method, a domain expert consultation method, and a Shap interpretable machine learning method; A training unit 203 is configured to train an extreme gradient boosting algorithm model based on the training data with the optimal AUC as a training direction to obtain an alternative model; A second construction unit 204 is used to construct an external independent validation data set for external validation of the model; A verification unit 205 is configured to verify the candidate model based on the external independent verification data set; The second screening unit 206 is used to select the candidate model with the best verification effect as the rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

[0062] like Figure 3 As shown, an embodiment of the present application also provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor. When the processor 320 executes the computer program 311, the steps of the above-mentioned method for screening patients with mild cognitive impairment based on a machine learning model are implemented.

[0063] Since the electronic device introduced in this embodiment is a device used to implement a mild cognitive impairment patient screening device based on a machine learning model in the embodiment of this application, based on the method introduced in the embodiment of this application, technical personnel in this field can understand the specific implementation of the electronic device of this embodiment and its various variations. Therefore, how the electronic device implements the method in the embodiment of this application will not be introduced in detail here. As long as the equipment used by technical personnel in this field to implement the method in the embodiment of this application falls within the scope of protection of this application.

[0064] During the specific implementation process, when the computer program 311 is executed by the processor, any implementation method of the embodiments corresponding to the first aspect can be implemented.

[0065] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing an electronic device according to an embodiment of the present application is shown.

[0066] It should be noted that Figure 4 The computer system 400 of the electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present application.

[0067] like Figure 4 As shown, computer system 400 includes a central processing unit (CPU) 401, which can perform various appropriate actions and processes, such as the methods described in the above embodiments, based on programs stored in read-only memory (ROM) 402 or programs loaded from storage 408 into random access memory (RAM) 403. RAM 403 also stores various programs and data required for system operation. CPU 401, ROM 402, and RAM 403 are interconnected via bus 404. An input / output (I / O) interface 405 is also connected to bus 404.

[0068] The following components are connected to the I / O interface 405: an input section 406 including a keyboard, mouse, and the like; an output section 407 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage section 408 including devices such as a hard disk; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card or a modem. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the I / O interface 405 as needed. Removable media 411, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 410 as needed, so that computer programs read from the media can be installed in the storage section 408 as needed.

[0069] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for executing the methods illustrated in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via the communication section 409 and / or installed from removable media 411. When executed by the central processing unit (CPU) 401, the computer program performs the various functions defined in the system of the present application.

[0070] It should be noted that the computer-readable medium described in the embodiments of this application may be a computer-readable signal medium or a computer-readable storage medium, or any combination thereof. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to, an electrical connection having one or more conductors, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium may include a data signal transmitted in baseband or as part of a carrier wave, which carries computer-readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0071] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. Among them, each box in the flowchart or block diagram can represent a module, program segment, or part of the code, and the above-mentioned module, program segment, or part of the code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0072] The units involved in the embodiments described in this application may be implemented by software or hardware, and the units described may also be set in a processor. In some cases, the names of these units do not constitute limitations on the units themselves.

[0073] As another aspect, the present application further provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the method for screening patients with mild cognitive impairment based on a machine learning model as described in the above embodiment.

[0074] As another aspect, the present application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments, or may exist independently without being incorporated into the electronic device. The computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device implements the method for screening patients with mild cognitive impairment based on a machine learning model as described in the above embodiments.

[0075] It should be noted that, although several modules or units of the device for action execution are mentioned in the above detailed description, this division is not mandatory. In fact, according to the embodiment of the application, the features and functions of two or more modules or units described above can be concretized in one module or unit. On the contrary, the features and functions of one module or unit described above can be further divided into multiple modules or units to be concretized.

[0076] Through the description of the above embodiments, it is easy for those skilled in the art to understand that the example embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solution according to the embodiments of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.) or on a network, and includes a number of instructions to enable a computing device (which can be a personal computer, a server, a touch terminal, or a network device, etc.) to execute the method according to the embodiments of the present application.

[0077] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of the present application that follow the general principles of the present application and include common knowledge or customary techniques in the art that are not disclosed in this application. It should be understood that this application is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from the scope thereof. The scope of this application is limited only by the appended claims.

Claims

1. A method for screening patients with mild cognitive impairment based on a machine learning model, characterized in that: include: Construct a cognitive health cohort of natural populations in the community and obtain cognitive-related assessment data of cohort study samples; Based on literature review, professional domain knowledge injection, domain expert consultation, and Shap interpretable machine learning, characteristic variables related to mild cognitive impairment are screened from the evaluation data as training data; Taking the best AUC as the training direction, training the extreme gradient boosting algorithm model based on the training data to obtain an alternative model; Construct an external independent validation dataset for external validation of the model; Validating the candidate model based on the external independent validation dataset; The alternative model with the best validation effect was selected as the rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

2. The method according to claim 1, characterized in that The construction of a community natural population cognitive health cohort and the acquisition of cognition-related assessment data of cohort study samples include: Population samples were recruited from natural population communities. Using a cohort research strategy as the starting point, comprehensive cognitive health assessment / diabetes comprehensive assessment was conducted on all subjects. Basic demographic information, cognitive and influencing factor assessment data, physical examination and biological specimen testing information of the sample population were systematically collected.

3. The method according to claim 1, characterized in that In the process of constructing a cognitive health cohort of the natural population in the community, a three-level quality control strategy was adopted, and quality control work was carried out at the three levels of preparation, execution, and maintenance.

4. The method according to claim 1, wherein The method based on literature review, professional domain knowledge injection, domain expert consultation and Shap interpretable machine learning is used to screen characteristic variables related to mild cognitive impairment from the evaluation data as training data, including: Based on the literature review method, we collected the influencing factors related to mild cognitive impairment from the literature related to mild cognitive impairment and formed a review list of influencing factors of mild cognitive impairment; Screening characteristic variables closely related to the occurrence of cognitive impairment from the evaluation data based on the review list of factors affecting mild cognitive impairment; An evaluation model is developed based on the review list of influencing factors of mild cognitive impairment, wherein the evaluation indicators of the evaluation model include: blood collection indicators, lifestyle characteristics, scale assessment characteristics, cognitive assessment indicators, previous chronic disease characteristics, and physical dimension indicators; Conducting pre-experimental verification of the evaluation model based on expert experience and evaluation scenarios; The validated assessment model was used to evaluate the characteristic variables closely related to the occurrence of cognitive impairment, and the assessment results were cleaned and sorted to obtain a set of candidate characteristic variables related to mild cognitive impairment. Perform category classification processing on the candidate feature variable set to obtain multiple feature subsets; The features in each feature subset are screened using domain expert consultation method; The filtered feature subsets are sequentially combined into single-mode and multi-mode combinations to obtain the total candidate feature variable set; Using the Shap interpretable machine learning method, with mild cognitive impairment as the label, the features in each feature subset in the total candidate feature variable set are ranked by their importance in terms of their contribution to label prediction, and the feature distribution results of multiple groups of feature importance rankings are obtained; The feature distribution results of each group of feature importance rankings are sorted out, and combined with expert experience, the top N feature variables in each feature subset are retained as training data, where N is an integer greater than or equal to 1.

5. The method according to claim 4, characterized in that The method of training an extreme gradient boosting algorithm model based on the training data with the optimal AUC as the training direction to obtain an alternative model includes: The optimal AUC is used as the training direction for model training based on each feature combination variable. The extreme gradient boosting algorithm model is trained based on each feature combination variable to obtain multiple preliminary models. Establish model screening criteria based on model performance; Based on the model screening conditions, a model that meets the requirements is selected from multiple preliminary models as an alternative model.

6. The method according to any one of claims 1 to 5, characterized in that The external independent validation dataset constructed for external validation of the model includes: Cognitive-related assessment data of apparently healthy middle-aged and elderly people were collected from communities that were not the source of model training data to form an external independent validation dataset.

7. The method according to claim 6, characterized in that The validating the candidate model based on the external independent validation dataset comprises: Clean and organize the external independent validation data set, and extract the cognitive screening assessment result label data required for external validation and the alternative model feature data entering the external validation phase; The alternative models are verified one by one based on the cognitive screening assessment result label data and the alternative model feature data.

8. A device for screening patients with mild cognitive impairment based on a machine learning model, characterized in that: include: The first construction unit is used to construct a cognitive health cohort of natural populations in the community and obtain cognitive-related evaluation data of cohort study samples; A first screening unit is configured to screen characteristic variables related to mild cognitive impairment from the evaluation data as training data based on a literature review method, a professional domain knowledge injection method, a domain expert consultation method, and a Shap interpretable machine learning method; A training unit, configured to train an extreme gradient boosting algorithm model based on the training data with the optimal AUC as a training direction to obtain an alternative model; The second construction unit is used to construct an external independent validation dataset for external validation of the model; a verification unit, configured to verify the candidate model based on the external independent verification data set; The second screening unit is used to select the alternative model with the best verification effect as a rapid screening model for mild cognitive impairment to screen patients with mild cognitive impairment.

9. A computer-readable storage medium, characterized in that The storage medium stores computer instructions, which, when executed on a computer, enable the computer to execute the method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: include: memory and processor; The memory is used to store computer instructions; The processor is configured to call the computer instructions stored in the memory so that the electronic device executes the method according to any one of claims 1 to 7.

Images