Cognitive disorder risk prediction method and device

Abstract

The invention provides a cognitive disorder risk prediction method and device. The method comprises the following steps: acquiring a feature value of at least one pre-selected user feature of an assessed person; inputting the feature value of the at least one pre-selected user feature of the assessed person into a cognitive disorder risk prediction model obtained by pre-training, and obtaining a cognitive disorder risk prediction result of the assessed person output by the cognitive disorder risk prediction model; wherein the pre-selected user features are screened out from a plurality of user features, and the contribution degree of the pre-selected user features to the cognitive disease risk prediction precision is greater than a preset threshold value. According to the method and the device, the cognitive disease risk prediction is carried out by utilizing the pre-screened user features with relatively high contribution degree to the cognitive disease risk prediction precision, so that a relatively accurate prediction result is realized by utilizing relatively few user features, the prediction efficiency is improved, and the prediction result has relatively good interpretability.

Description

Technical Field

[0001] The present application relates to the technical field of machine learning and data analysis and processing, and specifically to a method and device for predicting the risk of dementia. Background Art

[0002] As people age, their brain functions gradually deteriorate, and they may also develop some common cognitive diseases (neurodegenerative diseases) in the elderly, including Alzheimer's disease, Parkinson's disease, multiple system atrophy, etc. Dementia is one of the common chronic diseases in middle-aged and elderly people, with a long incubation period and difficult to reverse. Most of these diseases have an insidious onset, progressive aggravation, and great disability and lethality, which brings heavy life and economic burdens to patients and their families, and has become one of the most serious problems in today's society.

[0003] Chinese patent application CN115862855A discloses a method for risk assessment of Alzheimer's disease, which includes: obtaining patient condition data, including brain imaging data, psychological assessment data, language and medical record data; preprocessing brain imaging data, extracting features from the patient's condition data; constructing a disease risk assessment model based on the features; obtaining the target object's condition data, and inputting the disease risk assessment model to assess whether the target object has a disease risk. The disease risk assessment model proposed by the method uses a bilinear residual network architecture and a hybrid attention module to improve the assessment range of the assessment model, reduce the impact of the decrease in accuracy due to network deepening, and ensure the accuracy of the assessment data. At the same time, the use of multi-kernel functions further improves the accuracy of the assessment data, improves the generalization ability of the assessment model, and ensures the stability of the assessment accuracy. However, the method requires the acquisition of patient brain imaging data, and the collection of brain imaging data is relatively difficult, and there are problems such as high cost and low efficiency. In addition, the disease risk assessment model constructed by the method is relatively complex and poorly interpretable. Summary of the invention

[0004] At least one embodiment of the present application provides a method and device for predicting the risk of dementia, which can accurately and efficiently assess the risk of dementia and has good interpretability.

[0005] According to one aspect of the present application, at least one embodiment provides a method for predicting dementia risk, comprising:

[0006] Obtaining a feature value of at least one preselected user feature of the person being evaluated;

[0007] Inputting a feature value of at least one preselected user feature of the person being evaluated into a pre-trained dementia risk prediction model to obtain a dementia risk prediction result of the person being evaluated output by the dementia risk prediction model;

[0008] The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

[0009] Optionally, the multiple user characteristics include at least two of the following characteristics:

[0010] Brain health characteristics;

[0011] Basic personal information and lifestyle characteristics;

[0012] Personal physical examination and medical diagnosis characteristics.

[0013] Optionally, the brain health feature includes at least one of the following features: reaction ability, concentration, memory, cognitive ability level, and a comprehensive score corresponding to the brain health feature;

[0014] The basic personal information and lifestyle characteristics include at least one of the following characteristics: age, gender, body mass index (BMI), medical history, smoking habits, drinking habits, exercise habits, eating habits, and sleeping conditions;

[0015] The personal physical examination and medical diagnosis characteristics include at least one of the following characteristics: blood pressure, heart rate, height, weight, blood test items, urine test items, diseases, stage of illness, medication status, hospitalization status, and complications.

[0016] Optionally, the acquisition of the characteristic value of the brain health feature includes: obtaining the cerebral blood flow change value of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; or obtaining the task execution result of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; based on one or more of the reaction score, concentration score and memory score of the assessed person, obtaining the comprehensive score corresponding to the cognitive ability level or brain health feature of the assessed person;

[0017] The acquisition of the basic personal information and life habit characteristics includes: acquiring the data reported by the person being evaluated on a predefined questionnaire to obtain the characteristic values ​​of the basic personal information and life habit characteristics of the person being evaluated;

[0018] The acquisition of the personal physical examination and medical diagnosis characteristics includes: acquiring characteristic values ​​of the personal physical examination and medical diagnosis characteristics of the person being evaluated through a data interface with a physical examination center database, a hospital case database, and an insurance claim database.

[0019] Optionally, before obtaining the feature value of at least one preselected user feature of the evaluated person, the method further includes:

[0020] The dementia risk prediction model is trained using the pre-acquired training set and test set.

[0021] Optionally, using a pre-acquired training set to train the dementia risk prediction model includes:

[0022] Obtaining a training set and a test set, wherein each of the training set and the test set includes a plurality of user data, each of which includes feature values ​​of a plurality of user features of a user and a dementia risk assessment result of the user;

[0023] Based on the machine learning algorithm of the tree model, a first prediction model is constructed, and the first prediction model is trained using the training set to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the dementia risk prediction result of the user;

[0024] Using the test set to test the preliminary prediction model, to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model;

[0025] traversing each user feature in the test set, and when traversing each user feature, testing the preliminary prediction model using the test set without the feature value of the currently traversed user feature, obtaining a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determining the contribution of the currently traversed user feature to the dementia risk prediction accuracy according to the first value and the second value;

[0026] According to the contribution of each user feature to the prediction accuracy of dementia risk, screen out user features greater than a preset threshold as the pre-selected user features;

[0027] A second prediction model is constructed based on a machine learning algorithm of a tree model; user features in the training set are filtered, the pre-selected user features are retained, and the second prediction model is trained using the filtered training set to obtain the dementia risk prediction model.

[0028] Optionally, during the training process of the first prediction model and the second prediction model, the gradient boosting machine (GBM) algorithm is used to split the continuous numerical user features, and based on the splitting points, the continuous numerical user features are converted into discrete user features.

[0029] Optionally, the dementia risk prediction result includes the predicted onset time of dementia, the predicted onset risk and the corresponding cause of the onset.

[0030] According to another aspect of the present application, at least one embodiment provides a dementia risk prediction device, comprising:

[0031] A first acquisition module is used to acquire a feature value of at least one pre-selected user feature of the person being evaluated;

[0032] A prediction module, configured to input a feature value of at least one preselected user feature of the subject to be evaluated into a pre-trained dementia risk prediction model, and obtain a dementia risk prediction result of the subject to be evaluated output by the dementia risk prediction model;

[0033] The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

[0034] Optionally, the multiple user characteristics include at least two of the following characteristics:

[0035] Brain health characteristics;

[0036] Basic personal information and lifestyle characteristics;

[0037] Personal physical examination and medical diagnosis characteristics.

[0038] Optionally, the above device further comprises:

[0039] The training module is used to train the dementia risk prediction model using the pre-acquired training set and test set.

[0040] Optionally, the training module is further used to:

[0041] Obtaining a training set and a test set, wherein each of the training set and the test set includes a plurality of user data, each of which includes feature values ​​of a plurality of user features of a user and a dementia risk assessment result of the user;

[0042] Based on the machine learning algorithm of the tree model, a first prediction model is constructed, and the first prediction model is trained using the training set to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the dementia risk prediction result of the user;

[0043] Using the test set to test the preliminary prediction model, to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model;

[0044] traversing each user feature in the test set, and when traversing each user feature, testing the preliminary prediction model using the test set without the feature value of the currently traversed user feature, obtaining a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determining the contribution of the currently traversed user feature to the dementia risk prediction accuracy according to the first value and the second value;

[0045] According to the contribution of each user feature to the prediction accuracy of dementia risk, screen out user features greater than a preset threshold as the pre-selected user features;

[0046] A second prediction model is constructed based on a machine learning algorithm of a tree model; user features in the training set are filtered, the pre-selected user features are retained, and the second prediction model is trained using the filtered training set to obtain the dementia risk prediction model.

[0047] Optionally, the training module is also used to split the user features of continuous values ​​by using the gradient boosting machine (GBM) algorithm during the training process of the first prediction model and the second prediction model, and convert the user features of continuous values ​​into discrete user features based on the splitting points.

[0048] According to another aspect of the present application, at least one embodiment provides a computer-readable storage medium, on which a program is stored. When the program is executed by a processor, the steps of the method described above are implemented.

[0049] Compared with the prior art, the dementia risk prediction method and device provided in the embodiment of the present application can achieve more accurate prediction results using relatively fewer user features. In addition, due to the reduced types of user features, the embodiment of the present application can reduce the amount of calculation for model training and application, and improve prediction efficiency. In addition, the prediction results of the embodiment of the present application have good interpretability. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present application. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0051] Figure 1 A schematic diagram of a flow chart of a method for predicting dementia risk according to an embodiment of the present application;

[0052] Figure 2 A schematic diagram of a process flow of model training in an embodiment of the present application;

[0053] Figure 3 A schematic diagram of a structure of a dementia risk prediction model according to an embodiment of the present application;

[0054] Figure 4 A schematic diagram of the structure of a dementia risk prediction device according to an embodiment of the present application;

[0055] Figure 5 Schematic diagram of another structure of the dementia risk prediction device according to an embodiment of the present application. DETAILED DESCRIPTION

[0056] The exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the accompanying drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided in order to enable a more thorough understanding of the present application and to fully convey the scope of the present application to those skilled in the art.

[0057] It should be understood that the "one embodiment" or "an embodiment" mentioned throughout the specification means that the specific features, structures or characteristics related to the embodiment are included in at least one embodiment of the present application. Therefore, "in one embodiment" or "in one embodiment" appearing throughout the specification may not necessarily refer to the same embodiment. In addition, these specific features, structures or characteristics can be combined in one or more embodiments in any suitable manner. The terms "first", "second", etc. in the specification and claims of the present application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described here can be implemented in an order other than those illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. "And / or" in the specification and claims represents at least one of the connected objects.

[0058] In the various embodiments of the present application, it should be understood that the size of the serial numbers of the following processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0059] The following description provides examples and does not limit the scope, applicability, or configuration set forth in the claims. Changes may be made to the functions and arrangements of the elements discussed without departing from the spirit and scope of the present disclosure. Various examples may appropriately omit, replace, or add various procedures or components. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, features described with reference to certain examples may be combined in other examples.

[0060] Please refer to Figure 1 The method for predicting dementia risk provided in the embodiments of the present application includes:

[0061] Step 11, obtaining a feature value of at least one pre-selected user feature of the person being evaluated.

[0062] Here, the pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold. How to select user features will be described in detail later. The plurality of user features include at least two of the following features: (1) brain health features; (2) personal basic information and living habits features; (3) personal physical examination and medical diagnosis features.

[0063] More specifically, the brain health characteristics include but are not limited to at least one of the following characteristics:

[0064] (a) Responsiveness, which is used to characterize the user’s reaction speed;

[0065] (b) concentration, which is used to characterize the user's concentration level;

[0066] (c) memory, used to characterize memory ability;

[0067] (d) cognitive ability level, the certification ability level may be ability level data generated based on one or more of the above-mentioned reaction ability, concentration and memory;

[0068] (e) The comprehensive score corresponding to the brain health feature is a score value generated based on one or more of the above-mentioned reaction ability, concentration and memory, and the scoring system corresponds to the certification ability level. For example, the certification ability level is divided into multiple levels, and the comprehensive score corresponding to the brain health feature is divided into multiple continuous and non-overlapping score intervals, and each level corresponds to a score interval.

[0069] The characteristic value of the brain health feature can be obtained by automated brain health detection technology, which can obtain the brain forehead hemoglobin concentration + APP program score, or the artificial scale method can replace the brain impact detection method to improve efficiency and reduce costs. For example, the brain health index data of the predictor can be quickly obtained through an intelligent device or an artificial simple scale. Specifically, the score of the reaction force, concentration or memory of the person being evaluated can be obtained by obtaining the change value of the brain blood flow when the person being evaluated performs a predefined reaction force task, concentration task or memory task. For example, the blood flow data of the user's forehead is detected by a blood flow detection device when the user performs a related task to obtain the above score. The embodiment of the present application can also obtain the score of the reaction force, concentration or memory of the person being evaluated by obtaining the task execution result when the person being evaluated performs a predefined reaction force task, concentration task or memory task. Then, based on one or more of the scores of the reaction force, concentration, and memory of the person being evaluated, the comprehensive score corresponding to the cognitive ability level or brain health feature of the person being evaluated is obtained.

[0070] The personal basic information and living habits characteristics include but are not limited to at least one of the following characteristics:

[0071] (a) age;

[0072] (b) gender;

[0073] (c) Body Mass Index (BMI);

[0074] (d) past medical history;

[0075] (e) smoking habits, for example, the number of cigarettes smoked per unit time (e.g., per week, per day, etc.) or the duration of smoking;

[0076] (f) drinking habits, for example, the amount of alcohol consumed per unit time (e.g., per week, per day, etc.);

[0077] (g) exercise habits, such as the duration of exercise per unit time (e.g., per week, per day, etc.) or the calories consumed during exercise;

[0078] (h) dietary habits, for example, a diet containing light sugar, a diet containing moderate sugar, a diet containing heavy sugar, a diet containing light salt, a diet containing moderate salt, a diet containing heavy salt, etc. The specific standards for salt and sugar may be pre-set, for example, the amount of sugar / salt consumed daily may be divided into multiple intervals corresponding to light, moderate and heavy sugar / salt, respectively;

[0079] (i) Sleep status, which can be specifically expressed by indicators such as total sleep time per day and deep sleep time.

[0080] In the embodiment of the present application, the basic personal information and life habit characteristics can be obtained by filling out a questionnaire, that is, the filling data of the predefined questionnaire of the assessee is obtained to obtain the characteristic values ​​of the basic personal information and life habit characteristics of the assessee;

[0081] The personal physical examination and medical diagnosis characteristics include but are not limited to at least one of the following characteristics:

[0082] (a) Blood pressure;

[0083] (b) heart rate;

[0084] (c) height;

[0085] (d) Weight;

[0086] (e) blood test items, which may specifically include various blood test items;

[0087] (f) Urine test items, which may specifically include various urine test items;

[0088] (g) any illness;

[0089] (h) Period of disease;

[0090] (i) Medication status;

[0091] (j) Hospitalization status;

[0092] (k) Comorbid conditions.

[0093] The embodiment of the present application can obtain the characteristic values ​​of the personal physical examination and medical diagnosis characteristics of the person being evaluated through a data interface with a physical examination center database, a hospital case database, and an insurance claim database.

[0094] It can be seen that the user features include user features of continuous values ​​(such as height, age, blood pressure, etc.) and user features of discrete values ​​(such as gender, etc.). In addition, it should be noted that the above user features in the embodiment of the present application are only an example that can be used in the embodiment of the present application, and are not used to limit the present application. The present application can also delete the above user features or add other user features as needed.

[0095] Step 12: inputting the feature value of at least one pre-selected user feature of the person being evaluated into a pre-trained dementia risk prediction model to obtain a dementia risk prediction result of the person being evaluated output by the dementia risk prediction model.

[0096] Here, the dementia risk prediction model is a pre-trained model, which can be a Gradient Boosting Machine (GBM) model based on a tree model. The input data of the dementia risk prediction model is the characteristic value of the at least one pre-selected user feature, and the output data is the dementia risk prediction result, which can specifically include the predicted onset time of dementia (such as within the next 3 years), the predicted risk of onset (such as high risk, medium-high risk, medium risk, medium-low risk, low risk, etc.) and its corresponding cause of onset description. In addition, the percentage value corresponding to the predicted risk of onset can also be output. The cause of onset description can include a description of the indicators that lead to the risk of onset (such as a heavy salty diet, etc.), as well as a comparison with the standard value.

[0097] Through the above steps, the embodiment of the present application can use the pre-screened user features to predict the risk of dementia, and these screened user features are features that contribute more to the accuracy of dementia risk prediction, so the embodiment of the present application can use relatively few user features to achieve more accurate prediction results. In addition, due to the reduced types of user features, the embodiment of the present application can reduce the amount of calculation for model training and application, and improve prediction efficiency. In addition, since the screened user features are highly correlated with dementia, the prediction results of the embodiment of the present application have better interpretability.

[0098] Before step 11, the embodiment of the present application uses the pre-obtained training set and test set to train the dementia risk prediction model. The user feature screening and model training are described in detail below.

[0099] Please refer to Figure 2 The model training of the embodiment of the present application includes the following steps:

[0100] Step 21 : obtaining a training set and a test set, wherein both the training set and the test set include multiple sets of user data, and each set of user data includes feature values ​​of multiple user features of a user and a dementia risk assessment result of the user.

[0101] Here, the dementia risk assessment result in the user data may be obtained in advance. For example, the dementia risk assessment result of the user may be generated according to the user's dementia history as a label to guide the supervised training of the model.

[0102] Step 22, construct a first prediction model based on a machine learning algorithm of a tree model, and use the training set to train the first prediction model to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the user's dementia risk prediction result.

[0103] Here, based on the machine learning algorithm of the tree model, a first prediction model (such as a GBM model or a LightGBM model) is constructed. The structure of the model can be referred to Figure 3 , which is a tree structure. The first prediction model is trained using the training set, and the model obtained by the training is the preliminary prediction model. At this time, the preliminary prediction model is trained using all user features in the user data. The types of the user features can refer to the above description.

[0104] Step 23: Test the preliminary prediction model using the test set to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model.

[0105] Here, the preliminary prediction model trained in step 32 is tested using the test set to obtain a test result of the dementia risk prediction accuracy of the preliminary prediction model, that is, a first value.

[0106] Step 24, traverse each user feature in the test set, and when traversing each user feature, test the preliminary prediction model using the test set without the currently traversed user feature to obtain a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determine the contribution of the currently traversed user feature to the dementia risk prediction accuracy based on the first value and the second value.

[0107] Here, to avoid each user feature in the user data, when each user feature is traversed, the feature value of the user feature in the test set is deleted, for example, the feature value of the user feature is set to empty (such as null), so as to obtain an updated test set, and then, the preliminary prediction model trained in step 32 is tested using the updated test set to obtain the test result of the dementia risk prediction accuracy of the preliminary prediction model, that is, the second value. The first value is subtracted from the second value to obtain the contribution of the currently traversed user feature to the dementia risk prediction accuracy.

[0108] Step 25 , based on the contribution of each user feature to the prediction accuracy of dementia risk, select user features greater than a preset threshold as the pre-selected user features.

[0109] Here, a threshold is pre-set, and the size of the threshold can be flexibly set according to the model accuracy requirement. For example, when the model accuracy requirement is high, the threshold can be set smaller, and vice versa, when the model accuracy requirement is low, the threshold can be set larger. For example, based on the statistical confidence interval, the threshold can be set to 0.05. Of course, the embodiment of the present application can also be set to other values. Then, the user features whose contribution is greater than the threshold are screened out as pre-selected user features.

[0110] Step 26, constructing a second prediction model based on a machine learning algorithm of a tree model; filtering the user features in the training set, retaining the pre-selected user features, and using the filtered training set to train the second prediction model to obtain the dementia risk prediction model.

[0111] Here, the user features in the training set are filtered, user features whose contribution is less than or equal to the threshold are removed, and the pre-selected user features are retained to obtain an updated training set. Then, a new second prediction model is trained using the updated training set, and the model obtained by training is the dementia risk prediction model.

[0112] In the above steps 22 and 26, during the training process of the first prediction model and the second prediction model, the embodiment of the present application uses the GBM algorithm to split the user features of the continuous numerical values, and converts the user features of the continuous numerical values ​​into discrete user features based on the splitting points.

[0113] Table 1

[0114] User characteristics Minimum Maximum Scale age 30 75 15 BMI 22 30.5 3 Fasting blood sugar 3.9 6.1 1 hypertension 120 160 10 …… …… …… ……

[0115] Table 1 provides an example of splitting a user feature of continuous values. For example, for the age feature, the numerical range of the user feature is divided into multiple continuous and non-overlapping numerical intervals with a scale (interval) of 15 years, and each numerical interval is used as a discrete value of the user feature.

[0116] From the above, it can be seen that the embodiment of the present application realizes optimal discretization of features through a combination of GBM + customized user feature selection algorithm, removes features with smaller contributions, and improves model performance; and based on customized parameters, it is able to set feature selection thresholds, which is suitable for customers to flexibly choose between the complexity of high precision and the convenience of high availability for the model according to business needs and convenience.

[0117] Please refer to Figure 4 , a structure of a dementia risk prediction device provided in an embodiment of the present application, the dementia risk prediction device comprising:

[0118] A first acquisition module 41 is used to acquire a feature value of at least one pre-selected user feature of the evaluated person;

[0119] A prediction module 42, configured to input a feature value of at least one preselected user feature of the subject to be evaluated into a pre-trained dementia risk prediction model, and obtain a dementia risk prediction result of the subject to be evaluated output by the dementia risk prediction model;

[0120] The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

[0121] Optionally, the multiple user characteristics include at least two of the following characteristics:

[0122] Brain health characteristics;

[0123] Basic personal information and lifestyle characteristics;

[0124] Personal physical examination and medical diagnosis characteristics.

[0125] Optionally, the brain health feature includes at least one of the following features: reaction ability, concentration, memory, cognitive ability level, and a comprehensive score corresponding to the brain health feature;

[0126] The basic personal information and lifestyle characteristics include at least one of the following characteristics: age, gender, body mass index (BMI), medical history, smoking habits, drinking habits, exercise habits, eating habits, and sleeping conditions;

[0127] The personal physical examination and medical diagnosis characteristics include at least one of the following characteristics: blood pressure, heart rate, height, weight, blood test items, urine test items, diseases, stage of illness, medication status, hospitalization status, and complications.

[0128] Optionally, the acquisition of the characteristic value of the brain health feature includes: obtaining the cerebral blood flow change value of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; or obtaining the task execution result of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; based on one or more of the reaction score, concentration score and memory score of the assessed person, obtaining the comprehensive score corresponding to the cognitive ability level or brain health feature of the assessed person;

[0129] The acquisition of the basic personal information and life habit characteristics includes: acquiring the data reported by the person being evaluated on a predefined questionnaire to obtain the characteristic values ​​of the basic personal information and life habit characteristics of the person being evaluated;

[0130] The acquisition of the personal physical examination and medical diagnosis characteristics includes: acquiring characteristic values ​​of the personal physical examination and medical diagnosis characteristics of the person being evaluated through a data interface with a physical examination center database, a hospital case database, and an insurance claim database.

[0131] Optionally, the above device further includes:

[0132] The training module is used to train the dementia risk prediction model using the pre-acquired training set and test set.

[0133] Optionally, the training module is further used to:

[0134] Obtaining a training set and a test set, wherein each of the training set and the test set includes a plurality of user data, each of which includes feature values ​​of a plurality of user features of a user and a dementia risk assessment result of the user;

[0135] Based on the machine learning algorithm of the tree model, a first prediction model is constructed, and the first prediction model is trained using the training set to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the dementia risk prediction result of the user;

[0136] Using the test set to test the preliminary prediction model, to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model;

[0137] traversing each user feature in the test set, and when traversing each user feature, testing the preliminary prediction model using the test set without the feature value of the currently traversed user feature, obtaining a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determining the contribution of the currently traversed user feature to the dementia risk prediction accuracy according to the first value and the second value;

[0138] According to the contribution of each user feature to the prediction accuracy of dementia risk, screen out user features greater than a preset threshold as the pre-selected user features;

[0139] A second prediction model is constructed based on a machine learning algorithm of a tree model; user features in the training set are filtered, the pre-selected user features are retained, and the second prediction model is trained using the filtered training set to obtain the dementia risk prediction model.

[0140] Optionally, the training module is also used to split the user features of continuous values ​​by using the gradient boosting machine (GBM) algorithm during the training process of the first prediction model and the second prediction model, and convert the user features of continuous values ​​into discrete user features based on the splitting points.

[0141] Optionally, the dementia risk prediction result includes the predicted onset time of dementia, the predicted onset risk and the corresponding cause of the onset.

[0142] It should be noted that each system provided in the above embodiments is a device corresponding to the above dementia risk prediction method, and the implementation methods in the above embodiments are applicable to the embodiments of the device, and can also achieve the same technical effects. The above device provided in the embodiment of the present application can implement all the method steps implemented in the above method embodiment, and can achieve the same technical effects. The parts and beneficial effects that are the same as those in the method embodiment will not be specifically described here.

[0143] Please refer to Figure 5 , a schematic diagram of the structure of another dementia risk prediction device provided in an embodiment of the present application, the device includes: a processor 501, a transceiver 502, a memory 503, a user interface 504 and a bus interface.

[0144] In the embodiment of the present application, the device further includes: a program stored in the memory 503 and executable on the processor 501 .

[0145] The transceiver 502 is used to send and receive data under the control of the processor;

[0146] The processor 501 is configured to read the computer program in the memory and perform the following operations:

[0147] Obtaining a feature value of at least one preselected user feature of the person being evaluated;

[0148] Inputting a feature value of at least one preselected user feature of the person being evaluated into a pre-trained dementia risk prediction model to obtain a dementia risk prediction result of the person being evaluated output by the dementia risk prediction model;

[0149] The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

[0150] It is understandable that in the embodiment of the present application, when the computer program is executed by the processor 501, each process of the above-mentioned dementia risk prediction method embodiment can be implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0151] exist Figure 5In the embodiment, the bus architecture may include any number of interconnected buses and bridges, specifically linking together various circuits of one or more processors represented by processor 501 and memory represented by memory 503. The bus architecture may also link together various other circuits such as peripherals, voltage regulators, and power management circuits, which are well known in the art and are therefore not further described herein. The bus interface provides an interface. The transceiver 502 may be a plurality of components, namely, a transmitter and a receiver, providing a unit for communicating with various other devices on a transmission medium. For different user devices, the user interface 504 may also be an interface capable of externally and internally connecting required devices, and the connected devices include but are not limited to a keypad, a display, a speaker, a microphone, a joystick, and the like.

[0152] The processor 501 is responsible for managing the bus architecture and general processing, and the memory 503 can store data used by the processor 501 when performing operations.

[0153] It should be noted that the device in this embodiment is a device corresponding to the above-mentioned dementia risk prediction method, and the implementation methods in the above-mentioned embodiments are all applicable to the embodiments of the device, and can also achieve the same technical effects. In the device, the transceiver 502 and the memory 503, as well as the transceiver 502 and the processor 501 can be connected through the bus interface communication, and the function of the processor 501 can also be implemented by the transceiver 502, and the function of the transceiver 502 can also be implemented by the processor 501. It should be noted that the above-mentioned device provided in the embodiment of the present application can implement all the method steps implemented in the above-mentioned method embodiment, and can achieve the same technical effects, and the parts and beneficial effects that are the same as the method embodiment in this embodiment will not be specifically repeated here.

[0154] In some embodiments of the present application, a computer-readable storage medium is further provided, on which a program is stored, and when the program is executed by a processor, the following steps are implemented:

[0155] Obtaining a feature value of at least one preselected user feature of the person being evaluated;

[0156] Inputting a feature value of at least one preselected user feature of the person being evaluated into a pre-trained dementia risk prediction model to obtain a dementia risk prediction result of the person being evaluated output by the dementia risk prediction model;

[0157] The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

[0158] When the program is executed by a processor, it can implement all the implementation methods of the above-mentioned dementia risk prediction method and achieve the same technical effect. To avoid repetition, it will not be described here.

[0159] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0160] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0161] In the embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0162] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0163] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0164] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard drives, ROM, RAM, magnetic disks, or optical disks.

[0165] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for predicting the risk of dementia, characterized in that: include: Obtaining a feature value of at least one preselected user feature of the person being evaluated; Inputting a feature value of at least one preselected user feature of the person being evaluated into a pre-trained dementia risk prediction model to obtain a dementia risk prediction result of the person being evaluated output by the dementia risk prediction model; The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

2. The method according to claim 1, characterized in that The multiple user characteristics include at least two of the following characteristics: Brain health characteristics; Basic personal information and lifestyle characteristics; Personal physical examination and medical diagnosis characteristics.

3. The method according to claim 2, characterized in that The brain health characteristics include at least one of the following characteristics: reaction ability, concentration, memory, cognitive ability level, and a comprehensive score corresponding to the brain health characteristics; The basic personal information and lifestyle characteristics include at least one of the following characteristics: age, gender, body mass index (BMI), medical history, smoking habits, drinking habits, exercise habits, eating habits, and sleeping conditions; The personal physical examination and medical diagnosis characteristics include at least one of the following characteristics: blood pressure, heart rate, height, weight, blood test items, urine test items, diseases, stage of illness, medication status, hospitalization status, and complications.

4. The method according to claim 3, characterized in that Acquiring the characteristic value of the brain health characteristic includes: obtaining the cerebral blood flow change value of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; or obtaining the task execution result of the assessed person when performing a predefined reaction task, concentration task or memory task, thereby obtaining the reaction score, concentration score or memory score of the assessed person; obtaining the cognitive ability level of the assessed person or the comprehensive score corresponding to the brain health characteristic based on one or more of the reaction score, concentration score and memory score of the assessed person; The acquisition of the basic personal information and life habit characteristics includes: acquiring the data reported by the person being evaluated on a predefined questionnaire to obtain the characteristic values ​​of the basic personal information and life habit characteristics of the person being evaluated; The acquisition of the personal physical examination and medical diagnosis characteristics includes: acquiring characteristic values ​​of the personal physical examination and medical diagnosis characteristics of the person being evaluated through a data interface with a physical examination center database, a hospital case database, and an insurance claim database.

5. The method according to any one of claims 1 to 4, characterized in that: Before obtaining the feature value of at least one preselected user feature of the evaluated person, the method further includes: The dementia risk prediction model is trained using the pre-acquired training set and test set.

6. The method according to claim 5, characterized in that The dementia risk prediction model is trained using the pre-acquired training set, including: Obtaining a training set and a test set, wherein each of the training set and the test set includes a plurality of user data, each of which includes feature values ​​of a plurality of user features of a user and a dementia risk assessment result of the user; Based on the machine learning algorithm of the tree model, a first prediction model is constructed, and the first prediction model is trained using the training set to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the dementia risk prediction result of the user; Using the test set to test the preliminary prediction model, to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model; traversing each user feature in the test set, and when traversing each user feature, testing the preliminary prediction model using the test set without the feature value of the currently traversed user feature, obtaining a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determining the contribution of the currently traversed user feature to the dementia risk prediction accuracy according to the first value and the second value; According to the contribution of each user feature to the prediction accuracy of dementia risk, screen out user features greater than a preset threshold as the pre-selected user features; A second prediction model is constructed based on a machine learning algorithm of a tree model; user features in the training set are filtered, the pre-selected user features are retained, and the second prediction model is trained using the filtered training set to obtain the dementia risk prediction model.

7. The method according to claim 6, characterized in that During the training of the first prediction model and the second prediction model, the gradient boosting machine (GBM) algorithm is used to split the continuous numerical user features, and based on the splitting points, the continuous numerical user features are converted into discrete user features.

8. The method according to claim 1, characterized in that The dementia risk prediction result includes the predicted onset time of dementia, the predicted onset risk and the corresponding cause of the disease.

9. A dementia risk prediction device, characterized in that: include: A first acquisition module is used to acquire a feature value of at least one pre-selected user feature of the person being evaluated; A prediction module, configured to input a feature value of at least one preselected user feature of the subject to be evaluated into a pre-trained dementia risk prediction model, and obtain a dementia risk prediction result of the subject to be evaluated output by the dementia risk prediction model; The pre-selected user features are selected from a plurality of user features, and the contribution of the pre-selected user features to the accuracy of dementia risk prediction is greater than a preset threshold.

10. The device according to claim 9, characterized in that The multiple user characteristics include at least two of the following characteristics: Brain health characteristics; Basic personal information and lifestyle characteristics; Personal physical examination and medical diagnosis characteristics.

11. The device according to any one of claims 9 to 10, characterized in that Also includes: The training module is used to train the dementia risk prediction model using the pre-acquired training set and test set.

12. The device according to claim 11, characterized in that The training module is also used for: Obtaining a training set and a test set, wherein each of the training set and the test set includes a plurality of user data, each of which includes feature values ​​of a plurality of user features of a user and a dementia risk assessment result of the user; Based on the machine learning algorithm of the tree model, a first prediction model is constructed, and the first prediction model is trained using the training set to obtain a preliminary prediction model, wherein the input of the prediction model is the user data in the training set, and the output is the dementia risk prediction result of the user; Using the test set to test the preliminary prediction model, to obtain a first value of the dementia risk prediction accuracy of the preliminary prediction model; traversing each user feature in the test set, and when traversing each user feature, testing the preliminary prediction model using the test set without the feature value of the currently traversed user feature, obtaining a second value of the dementia risk prediction accuracy of the preliminary prediction model, and determining the contribution of the currently traversed user feature to the dementia risk prediction accuracy according to the first value and the second value; According to the contribution of each user feature to the prediction accuracy of dementia risk, screen out user features greater than a preset threshold as the pre-selected user features; A second prediction model is constructed based on a machine learning algorithm of a tree model; user features in the training set are filtered, the pre-selected user features are retained, and the second prediction model is trained using the filtered training set to obtain the dementia risk prediction model.

13. The device according to claim 12, characterized in that The training module is also used to split the user features of continuous values ​​by using the gradient boosting machine (GBM) algorithm during the training process of the first prediction model and the second prediction model, and convert the user features of continuous values ​​into discrete user features based on the splitting points.

Images