A system and method for predicting Alzheimer's disease risk based on continuous data

Abstract

The present invention provides a continuous data-based Alzheimer's disease risk prediction system and method, comprising: a data acquisition module for collecting multi-dimensional continuous data from an individual; a data analysis module for analyzing the multi-dimensional continuous data, collecting data features from the multi-dimensional continuous data, and outputting individual posture features based on the data features; and a prediction module for determining, based on the posture features, the change points in the individual's behavioral path within a target time series and predicting Alzheimer's disease risk based on these change points. This ensures the accuracy and reliability of Alzheimer's disease risk prediction, greatly facilitating disease treatment.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a system and method for predicting Alzheimer's disease risk based on continuous data. Background Art

[0002] Currently, Alzheimer's disease is a neurodegenerative disease that seriously affects human health. Early diagnosis and risk prediction are crucial. Under current technological background, the acquisition and utilization of continuous data has become key, and continuous monitoring can be achieved through various advanced monitoring devices.

[0003] However, existing technologies cannot accurately and effectively analyze data, making it impossible to identify and analyze subtle changes in the data. This not only reduces the accuracy of Alzheimer's disease risk prediction, but also fails to guarantee the reliability of the prediction results.

[0004] Therefore, in order to overcome the above technical problems, the present invention provides an Alzheimer's disease risk prediction system and method based on continuous data. Summary of the Invention

[0005] The present invention provides an Alzheimer's disease risk prediction system and method based on continuous data. The system collects multi-dimensional continuous data from an individual to provide reliable data support for Alzheimer's disease risk prediction. Secondly, the collected multi-dimensional continuous data is analyzed to accurately and effectively extract data features from the multi-dimensional continuous data, facilitating the determination of individual posture features. Finally, the determined posture features are parsed to accurately determine the individual's behavioral path change points, effectively determining the individual's movement and behavior within a target time series. Ultimately, accurate and reliable prediction of Alzheimer's disease risk is achieved based on the behavioral path change points, ensuring the accuracy and reliability of Alzheimer's disease risk prediction and greatly facilitating disease treatment.

[0006] The present invention provides an Alzheimer's disease risk prediction system based on continuous data, comprising:

[0007] Data collection module, used to collect multi-dimensional continuous data of individuals;

[0008] The data analysis module is used to analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features;

[0009] The prediction module is used to determine the behavioral path change points of an individual in the target time series based on posture features, and to predict the risk of Alzheimer's disease based on the behavioral path change points.

[0010] Preferably, a system for predicting Alzheimer's disease risk based on continuous data, the data acquisition module comprises:

[0011] A data collection preparation unit, configured to determine the collection dimensions of individual continuous data based on the collection requirements, and to determine the database corresponding to each collection dimension;

[0012] The mechanism configuration unit is used to configure data collection mechanisms for different databases and connect the data collection mechanisms with the databases. At the same time, based on the collection requirements, it determines the collection time sequence for continuous data under each collection dimension and configures the parameters of the data collection mechanism based on the collection time sequence;

[0013] The data collection unit is used to control the data collection mechanism based on the parameter configuration results to collect data from the databases under the corresponding collection dimensions, and to summarize and align the data collection results of each database based on the spatiotemporal characteristics to obtain individual multi-dimensional continuous data.

[0014] Preferably, a system for predicting Alzheimer's disease risk based on continuous data, the data acquisition unit comprises:

[0015] The data retrieval subunit is used to obtain the obtained multi-dimensional continuous data and visualize the multi-dimensional continuous data in a preset coordinate system in turn;

[0016] Data preprocessing subunit, used to:

[0017] Determine the changing trend of continuous data under each collection dimension based on the visual display results, and identify isolated sample points based on the changing trend;

[0018] Extract the upper and lower sample values ​​of the isolated sample point, and modify the target value of the isolated sample point based on the mean of the upper and lower sample values;

[0019] Obtaining sample characterization requirements under each collection dimension, and determining a standardized interval for continuous data under each collection dimension based on the sample characterization requirements;

[0020] The corrected continuous data of each collected dimension are standardized based on the standardized interval.

[0021] Preferably, a data analysis module of an Alzheimer's disease risk prediction system based on continuous data includes:

[0022] A dimension identification acquisition unit is used to obtain the dimension identification of the multi-dimensional continuous data, and retrieve the continuous sample data of each dimension in a normal state from a preset database according to the dimension identification of the multi-dimensional continuous data;

[0023] Target transformation curve construction unit, used for:

[0024] Reading continuous sample data of each dimension in a normal state, and performing a first mapping on the continuous sample data in a preset rectangular coordinate system according to a time sequence;

[0025] Obtain target transformation curves of various dimensions according to the first mapping result;

[0026] A curve analysis unit, configured to analyze the target transformation curve and obtain a reference interval of the target transformation curve;

[0027] a data feature determination unit, configured to display the continuous data in a rectangular coordinate system of a corresponding dimension, compare the displayed result with a reference interval, and determine the data feature of the multi-dimensional continuous data based on the comparison result;

[0028] The posture feature determination unit is used to determine the posture feature of the individual according to the data feature.

[0029] Preferably, an Alzheimer's disease risk prediction system based on continuous data, the curve analysis unit, comprises:

[0030] A second-order derivative operation subunit is used to read the target transformation curve, perform a second-order derivative operation on the target transformation curve, and determine the concave and convex points of the target transformation curve according to the second-order derivative operation result of the target transformation curve;

[0031] The first baseline determination subunit is configured to:

[0032] Marking the concave and convex points on the target transformation curve, calculating the first mean of each convex point, and determining the first reference value of the corresponding dimension according to the first mean of each convex point;

[0033] Performing a second mapping on the first reference value in the target transformation curve to obtain a first reference line;

[0034] The second baseline determination subunit is configured to:

[0035] Calculating the second mean of each concave point, and determining the second reference value of the corresponding dimension according to the second mean of each concave point;

[0036] Performing a third mapping of the second reference value in the target transformation curve to obtain a second reference line;

[0037] The reference interval determination subunit is configured to determine a reference interval of a target transformation interval according to the first reference line and the second reference line.

[0038] Preferably, a posture feature determination unit of an Alzheimer's disease risk prediction system based on continuous data includes:

[0039] When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the first posture feature is output;

[0040] When the data feature shows that the continuous data is greater than the reference interval in the rectangular coordinate system of the corresponding dimension, the second posture feature is output;

[0041] When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the third posture feature is output.

[0042] Preferably, an Alzheimer's disease risk prediction system based on continuous data, the prediction module comprises:

[0043] Data Access Unit, used to:

[0044] Based on the dimensional information of multi-dimensional continuous data, multi-dimensional historical data is retrieved from the database, and risk assessment factors are determined based on the preset assessment knowledge system of Alzheimer's disease.

[0045] Extracting data representations of multi-dimensional historical data, and determining the correlation index between the multi-dimensional historical data and risk assessment factors based on the data representations, and determining the multi-dimensional historical data with a correlation index greater than a preset threshold as a strongly correlated parameter;

[0046] At the same time, based on the preset evaluation knowledge system, the multi-dimensional topological structure between the strongly correlated parameters corresponding to the multi-dimensional historical data is determined, and the correlation attributes between the strongly correlated parameters are obtained based on the multi-dimensional topological structure;

[0047] Model training unit, used for:

[0048] Select a model framework from the model library based on the evaluation requirements and extract the model parameters of the model framework;

[0049] Iteratively train the model parameters based on strongly correlated parameters and correlated attributes, and monitor the training values ​​of each model indicator in real time during the iterative training process. When the training values ​​meet the preset requirements, terminate the iterative training to obtain a risk prediction model.

[0050] Risk prediction unit, used to:

[0051] Based on the time development sequence, the posture features corresponding to the multi-dimensional continuous data are separated into sequences to obtain the action parameters of each dimension of the individual continuous data at each moment, and the action parameters of adjacent moments are compared and statistically summarized to obtain the change points of the individual's behavior path under the target time sequence;

[0052] Based on the risk prediction model, risk analysis is performed on the behavioral path change points to obtain the risk probability of the individual under each dimension of continuous data. Based on the target weight of each dimension of continuous data, the risk probabilities corresponding to multiple dimensions are weighted averaged to obtain the individual's Alzheimer's risk prediction value;

[0053] At the same time, the individual's Alzheimer's disease pathological representation is determined based on the risk probability and target weight under each dimension of continuous data, and the Alzheimer's disease risk prediction value and the pathological representation are associated and bound.

[0054] Preferably, an Alzheimer's disease risk prediction system based on continuous data, the risk prediction unit, comprises:

[0055] a result acquisition subunit, configured to obtain the association binding result between the obtained Alzheimer's disease risk prediction value and the pathological representation, and at the same time, perform hierarchical matching between the Alzheimer's disease risk prediction value and a preset risk level comparison table, and obtain the Alzheimer's disease risk level based on the hierarchical matching result;

[0056] Record subunit, used to:

[0057] Associating the Alzheimer's disease risk level with the result subordinate tag, and determining the record items and the record representation of each record item based on the subordinate tag;

[0058] A record table is constructed based on the record items and the record representations, and the association binding results of the Alzheimer's disease risk level, the Alzheimer's disease risk prediction value and the pathological representation are recorded and stored in the record table.

[0059] Preferably, an Alzheimer's disease risk prediction system based on continuous data, the prediction module comprises:

[0060] A data reading unit is used to read the individual's multi-dimensional continuous data and determine the baseline data value of Alzheimer's disease in each dimension. At the same time, it reads the risk probability value of each dimension of continuous data based on Alzheimer's disease risk prediction;

[0061] a first calculation unit, configured to calculate the accuracy of the result of the Alzheimer's disease risk prediction based on the individual's multi-dimensional continuous data, the baseline data value of the occurrence of Alzheimer's disease in each dimension, and the risk probability value of the Alzheimer's disease risk prediction based on the continuous data in each dimension;

[0062] ;

[0063] in, represents the accuracy of the results of Alzheimer's disease risk prediction; Indicates the dimension sequence value; Indicates the total number of dimensions; Indicates the Continuous data in 3 dimensions; Indicates the The baseline data value of Alzheimer's disease in each dimension; Represents a constant, with a value of 2; Indicates the The influence coefficient corresponding to each dimension is (0.1, 0.2); Indicates the Risk probability value of continuous data based on Alzheimer's disease risk prediction in each dimension;

[0064] The second computing unit is configured to:

[0065] Obtain a total number of predictions for Alzheimer's disease risk prediction, and determine the number of accurate predictions and the number of incorrect predictions among the total number of predictions;

[0066] The prediction efficiency of Alzheimer's disease risk prediction is calculated based on the number of accurate predictions, the number of incorrect predictions, and the accuracy rate of Alzheimer's disease risk prediction results;

[0067] ;

[0068] in, represents the predictive efficiency for Alzheimer's disease risk prediction; Indicates the number of accurate predictions; Indicates the number of wrong predictions; represents the total number of predictions for Alzheimer's disease risk;

[0069] Qualification judgment unit, used for:

[0070] obtaining a prediction efficiency threshold, and comparing the prediction efficiency of the Alzheimer's disease risk prediction with the prediction efficiency threshold to determine whether the prediction of Alzheimer's disease is qualified;

[0071] If the prediction efficiency of the risk prediction of Alzheimer's disease is equal to or greater than the prediction efficiency threshold, then the prediction of Alzheimer's disease is determined to be qualified;

[0072] Otherwise, it is judged to be unqualified for the prediction of Alzheimer's disease;

[0073] The alarm unit is used for performing an alarm operation when it is determined that the prediction of Alzheimer's disease is unqualified.

[0074] The present invention provides a method for predicting Alzheimer's disease risk based on continuous data, comprising:

[0075] Step 1: Collect multi-dimensional continuous data of individuals;

[0076] Step 2: Analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features;

[0077] Step 3: Determine the user's behavior path change points in the target time series based on the posture features, and predict the risk of Alzheimer's disease based on the behavior path change points.

[0078] Compared with the prior art, the present invention has the following beneficial effects:

[0079] 1. By collecting multi-dimensional continuous data of individuals, reliable data support is provided for Alzheimer's disease risk prediction. Secondly, the collected multi-dimensional continuous data is analyzed to accurately and effectively extract the data features of the multi-dimensional continuous data, which facilitates the determination of individual posture characteristics. Finally, the determined posture characteristics are parsed to accurately determine the change points of the individual's behavior path, so as to effectively determine the individual's action behavior in the target time series. Ultimately, accurate and reliable prediction of Alzheimer's disease risk is achieved based on the change points of the behavior path, ensuring the accuracy and reliability of Alzheimer's disease risk prediction and providing great convenience for disease treatment.

[0080] 2. By retrieving the corresponding multi-dimensional historical data from the database based on the dimensional information and analyzing the multi-dimensional historical data, it is possible to screen out strongly correlated parameters that are highly correlated with risk assessment factors from the multi-dimensional historical data, thereby providing reliable training samples for risk prediction. Secondly, by determining the correlation attributes between the strongly correlated parameters, the selected model framework is accurately and effectively trained according to the strongly correlated parameters and the corresponding correlation attributes, ensuring the reliability of the final risk prediction model. Finally, the individual posture features are sequenced and the individual's behavior path change points are accurately and effectively locked. The obtained behavior path change points are input into the risk prediction model for risk prediction, and finally the Alzheimer's disease risk prediction value and pathological characterization are effectively determined, ensuring the accuracy and reliability of Alzheimer's disease risk prediction.

[0081] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be achieved and obtained through the structures specifically pointed out in this application document.

[0082] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0083] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0084] Figure 1 This is a structural diagram of an Alzheimer's disease risk prediction system based on continuous data in an embodiment of the present invention;

[0085] Figure 2 This is a structural diagram of a data acquisition module in an Alzheimer's disease risk prediction system based on continuous data in an embodiment of the present invention;

[0086] Figure 3 4 is a flow chart of a method for predicting Alzheimer's disease risk based on continuous data in an embodiment of the present invention. DETAILED DESCRIPTION

[0087] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0088] Example 1:

[0089] This embodiment provides an Alzheimer's disease risk prediction system based on continuous data, such as Figure 1 As shown, including:

[0090] Data collection module, used to collect multi-dimensional continuous data of individuals;

[0091] The data analysis module is used to analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features;

[0092] The prediction module is used to determine the behavioral path change points of an individual in the target time series based on posture features, and to predict the risk of Alzheimer's disease based on the behavioral path change points.

[0093] In this embodiment, the individual may be a user who needs to perform Alzheimer's disease risk prediction.

[0094] In this embodiment, the multi-dimensional continuous data may be different types of behavioral data corresponding to the user within a certain period of time, including physiological indicator data, behavioral data, cognitive test data, etc.

[0095] In this embodiment, the data features may be the changing trends and value changes of the multi-dimensional continuous data obtained after parsing the multi-dimensional continuous data, thereby achieving accurate and effective determination of the individual's posture features.

[0096] In this embodiment, the posture feature may be a behavioral representation of an individual under different types of continuous data, including facial expressions and language behaviors.

[0097] In this embodiment, the target time series may be a monitoring time period corresponding to multi-dimensional continuous data, the purpose of which is to split the posture features so as to determine the change points of the individual's behavior path.

[0098] In this embodiment, the behavior path change point may be a movement behavior change trajectory or change feature generated by the individual's posture characteristics developing over time.

[0099] The working principle and beneficial effects of the above technical solution are as follows: by collecting multi-dimensional continuous data of an individual, reliable data support is provided for Alzheimer's disease risk prediction; secondly, the collected multi-dimensional continuous data is analyzed to accurately and effectively extract the data features of the multi-dimensional continuous data, which facilitates the determination of the individual's posture characteristics; finally, the determined posture characteristics are parsed to accurately determine the change points of the individual's behavior path, thereby effectively determining the individual's action behavior in the target time series, and ultimately achieving accurate and reliable prediction of Alzheimer's disease risk based on the change points of the behavior path, thereby ensuring the accuracy and reliability of Alzheimer's disease risk prediction and greatly facilitating disease treatment.

[0100] Example 2:

[0101] Based on Example 1, this example provides an Alzheimer's disease risk prediction system based on continuous data, such as Figure 2 As shown, the data acquisition module includes:

[0102] A data collection preparation unit, configured to determine the collection dimensions of individual continuous data based on the collection requirements, and to determine the database corresponding to each collection dimension;

[0103] The mechanism configuration unit is used to configure data collection mechanisms for different databases and connect the data collection mechanisms with the databases. At the same time, based on the collection requirements, it determines the collection time sequence for continuous data under each collection dimension and configures the parameters of the data collection mechanism based on the collection time sequence;

[0104] The data collection unit is used to control the data collection mechanism based on the parameter configuration results to collect data from the databases under the corresponding collection dimensions, and to summarize and align the data collection results of each database based on the spatiotemporal characteristics to obtain individual multi-dimensional continuous data.

[0105] In this embodiment, the collection requirements are set in advance and are used to characterize the types of data that need to be collected and the collection standards that need to be obtained when collecting data.

[0106] In this embodiment, the collection dimension may be the type of data that is ultimately determined to be required to be collected.

[0107] In this embodiment, the data collection mechanism is a strategy for retrieving data from a database, and each database corresponds to a data collection mechanism.

[0108] In this embodiment, the collection time series may represent the time required for data collection, for example, it may be multi-dimensional continuous data collected from an individual within a week.

[0109] In this embodiment, the spatiotemporal feature may be to associate data collection results at the same time and space in order to ensure the integrity of data at the same dimension.

[0110] The working principle and beneficial effects of the above technical solution are: by parsing the collection requirements, the collection dimensions and the database corresponding to each collection dimension can be accurately and effectively determined; secondly, a corresponding data collection mechanism is configured for each database, and the data collection mechanism is configured according to the collection time series determined by the collection requirements, ensuring that the data collection mechanism can effectively collect the data in the database; finally, the collected data is summarized and aligned to achieve effective collection and acquisition of multi-dimensional continuous data, providing reliable data support for risk prediction.

[0111] Example 3:

[0112] Based on Example 2, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the data acquisition unit includes:

[0113] The data retrieval subunit is used to obtain the obtained multi-dimensional continuous data and visualize the multi-dimensional continuous data in a preset coordinate system in turn;

[0114] Data preprocessing subunit, used to:

[0115] Determine the changing trend of continuous data under each collection dimension based on the visual display results, and identify isolated sample points based on the changing trend;

[0116] Extract the upper and lower sample values ​​of the isolated sample point, and modify the target value of the isolated sample point based on the mean of the upper and lower sample values;

[0117] Obtaining sample characterization requirements under each collection dimension, and determining a standardized interval for continuous data under each collection dimension based on the sample characterization requirements;

[0118] The corrected continuous data of each collected dimension are standardized based on the standardized interval.

[0119] In this embodiment, the preset coordinate system is constructed in advance and is used to visualize the collected multi-dimensional continuous data, thereby facilitating pre-processing operations on the collected multi-dimensional continuous data.

[0120] In this embodiment, isolated sample points may be data samples that deviate from the overall change trend in the continuous data under each acquisition dimension.

[0121] In this embodiment, the upper and lower sample values ​​may be specific values ​​of two data adjacent to the isolated sample point.

[0122] In this embodiment, the sample characterization requirement may be a normalization standard for characterizing samples under each acquisition dimension, wherein the normalization interval is the parameter limit that is finally determined to be required for normalization, for example, the data may be normalized to the interval [0, 1].

[0123] The working principle and beneficial effects of the above technical solution are: by preprocessing the collected multi-dimensional continuous data, the accuracy, reliability and format uniformity of the multi-dimensional continuous data finally obtained are ensured, thereby providing reliable data information for Alzheimer's disease risk prediction.

[0124] Example 4:

[0125] Based on Example 1, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the data analysis module includes:

[0126] A dimension identification acquisition unit is used to obtain the dimension identification of the multi-dimensional continuous data, and retrieve the continuous sample data of each dimension in a normal state from a preset database according to the dimension identification of the multi-dimensional continuous data;

[0127] Target transformation curve construction unit, used for:

[0128] Reading continuous sample data of each dimension in a normal state, and performing a first mapping on the continuous sample data in a preset rectangular coordinate system according to a time sequence;

[0129] Obtain target transformation curves of various dimensions according to the first mapping result;

[0130] A curve analysis unit, configured to analyze the target transformation curve and obtain a reference interval of the target transformation curve;

[0131] a data feature determination unit, configured to display the continuous data in a rectangular coordinate system of a corresponding dimension, compare the displayed result with a reference interval, and determine the data feature of the multi-dimensional continuous data based on the comparison result;

[0132] The posture feature determination unit is used to determine the posture feature of the individual according to the data feature.

[0133] In this embodiment, the preset database may be set in advance and used to store continuous sample data of each dimension under normal conditions.

[0134] In this embodiment, the first mapping may be used to describe the changing state of continuous sample data in a rectangular coordinate system.

[0135] In this embodiment, the reference interval may be used to measure data features of continuous data.

[0136] The working principle and beneficial effects of the above technical solution are: by obtaining the dimension identification of multi-dimensional continuous data, the continuous sample data of each dimension in the normal state can be effectively and accurately read in the preset database, and the target change curve can be constructed through the first mapping, and the reference interval can be effectively determined through the analysis of the target change curve, and the data features of the multi-dimensional continuous data can be effectively extracted through the reference interval, that is, based on the size and other characteristics of the multi-dimensional continuous data of the reference interval, the effectiveness of obtaining the data features is further improved, thereby laying the foundation for determining the posture features of the individual.

[0137] Example 5:

[0138] Based on Example 4, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the curve analysis unit comprises:

[0139] A second-order derivative operation subunit is used to read the target transformation curve, perform a second-order derivative operation on the target transformation curve, and determine the concave and convex points of the target transformation curve according to the second-order derivative operation result of the target transformation curve;

[0140] The first baseline determination subunit is configured to:

[0141] Marking the concave and convex points on the target transformation curve, calculating the first mean of each convex point, and determining the first reference value of the corresponding dimension according to the first mean of each convex point;

[0142] Performing a second mapping on the first reference value in the target transformation curve to obtain a first reference line;

[0143] The second baseline determination subunit is configured to:

[0144] Calculating the second mean of each concave point, and determining the second reference value of the corresponding dimension according to the second mean of each concave point;

[0145] Performing a third mapping of the second reference value in the target transformation curve to obtain a second reference line;

[0146] The reference interval determination subunit is configured to determine a reference interval of a target transformation interval according to the first reference line and the second reference line.

[0147] In this embodiment, performing a second-order derivative operation on the target change curve can effectively determine the concave and convex point values ​​of the target transformation curve, thereby ensuring the effectiveness of marking the concave and convex points on the target transformation curve.

[0148] In this embodiment, the first reference line can be positioned on the y-axis of the rectangular coordinate system based on the first reference value, and a straight line parallel to the x-axis can be used as the first reference line.

[0149] In this embodiment, the second reference line can be positioned on the y-axis of the rectangular coordinate system based on the second reference value, and a straight line parallel to the y-axis can be used as the second reference line.

[0150] In this embodiment, the reference interval may be the difference between the area between the first reference line and the x-axis and the area between the second reference line and the x-axis.

[0151] The working principle and beneficial effects of the above technical solution are: by performing a second-order derivative operation on the target transformation curve, the position of the concave and convex points on the target transformation curve can be effectively and accurately determined, thereby providing an accurate data basis for determining the first reference value and the second reference value, thereby effectively determining the reference interval between the first reference line and the second reference line, ensuring the rationality and effectiveness of the reference interval value range, thereby effectively providing a guarantee for the subsequent acquisition of data features of continuous data.

[0152] Example 6:

[0153] Based on Example 4, this embodiment provides an Alzheimer's disease risk prediction system based on continuous data, wherein the posture feature determination unit includes:

[0154] When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the first posture feature is output;

[0155] When the data feature shows that the continuous data is greater than the reference interval in the rectangular coordinate system of the corresponding dimension, the second posture feature is output;

[0156] When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the third posture feature is output.

[0157] In this embodiment, the first posture feature may be an individual posture performance state corresponding to the continuous data within a reference interval, wherein the posture feature within the reference interval is a posture feature obtained based on actual experience and experiments.

[0158] In this embodiment, the second posture feature may be the individual posture performance state corresponding to the continuous data when the continuous data is greater than the reference interval, wherein the posture feature greater than the reference interval is a posture feature obtained based on actual experience and experiments.

[0159] In this embodiment, the third posture feature may be the individual posture performance state corresponding to the continuous data when the continuous data is smaller than the reference interval, wherein the posture feature smaller than the reference interval is a posture feature obtained based on actual experience and experiments.

[0160] The working principle and beneficial effect of the above technical solution are: effectively measuring the posture characteristics of continuous data in different value ranges in the data features through the reference interval, thereby effectively ensuring the accuracy and comprehensiveness of the posture characteristics obtained.

[0161] Example 7:

[0162] Based on Example 1, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the prediction module includes:

[0163] Data Access Unit, used to:

[0164] Based on the dimensional information of multi-dimensional continuous data, multi-dimensional historical data is retrieved from the database, and risk assessment factors are determined based on the preset assessment knowledge system of Alzheimer's disease.

[0165] Extracting data representations of multi-dimensional historical data, and determining the correlation index between the multi-dimensional historical data and risk assessment factors based on the data representations, and determining the multi-dimensional historical data with a correlation index greater than a preset threshold as a strongly correlated parameter;

[0166] At the same time, based on the preset evaluation knowledge system, the multi-dimensional topological structure between the strongly correlated parameters corresponding to the multi-dimensional historical data is determined, and the correlation attributes between the strongly correlated parameters are obtained based on the multi-dimensional topological structure;

[0167] Model training unit, used for:

[0168] Select a model framework from the model library based on the evaluation requirements and extract the model parameters of the model framework;

[0169] Iteratively train the model parameters based on strongly correlated parameters and correlated attributes, and monitor the training values ​​of each model indicator in real time during the iterative training process. When the training values ​​meet the preset requirements, terminate the iterative training to obtain a risk prediction model.

[0170] Risk prediction unit, used to:

[0171] Based on the time development sequence, the posture features corresponding to the multi-dimensional continuous data are separated into sequences to obtain the action parameters of each dimension of the individual continuous data at each moment, and the action parameters of adjacent moments are compared and statistically summarized to obtain the change points of the individual's behavior path under the target time sequence;

[0172] Based on the risk prediction model, risk analysis is performed on the behavioral path change points to obtain the risk probability of the individual under each dimension of continuous data. Based on the target weight of each dimension of continuous data, the risk probabilities corresponding to multiple dimensions are weighted averaged to obtain the individual's Alzheimer's risk prediction value;

[0173] At the same time, the individual's Alzheimer's disease pathological representation is determined based on the risk probability and target weight under each dimension of continuous data, and the Alzheimer's disease risk prediction value and the pathological representation are associated and bound.

[0174] In this embodiment, the dimension information refers to the data type corresponding to the multi-dimensional continuous data.

[0175] In this embodiment, the multi-dimensional historical data may be the symptom data corresponding to Alzheimer's disease that is known and effectively recorded and retrieved from the corresponding database according to the data type.

[0176] In this embodiment, the preset evaluation knowledge system is set in advance and is used to characterize all data and the relationships between the data that may be involved in Alzheimer's disease risk prediction.

[0177] In this embodiment, the risk assessment factors may be assessment indicators involved in Alzheimer's disease risk assessment, such as the individual's language orderliness and motor agility.

[0178] In this embodiment, the data representation may be data characteristics of the multi-dimensional historical data, including the data composition of the multi-dimensional historical data.

[0179] In this embodiment, the correlation index may represent the degree of correlation between the multi-dimensional historical data and the risk assessment factors, and a larger value indicates a greater degree of correlation between the two.

[0180] In this embodiment, the preset threshold is set in advance and can be adjusted.

[0181] In this embodiment, the strongly correlated parameter may be data in the multi-dimensional historical data that is highly correlated with the risk assessment factors, that is, key data for directly performing risk prediction.

[0182] In this embodiment, the multidimensional topological structure can be used to characterize the interaction relationship between strongly correlated parameters and to characterize the interconnection relationship between different strongly correlated parameters, wherein the association attributes are determined based on the multidimensional topological structure and are used to characterize the association relationship between different strongly correlated parameters.

[0183] In this embodiment, the evaluation requirements are set in advance.

[0184] In this embodiment, the model framework may be a basic model suitable for risk assessment selected from a model library according to assessment requirements and requiring further training.

[0185] In this embodiment, the model indicator can be a parameter used to measure whether the model training meets the training requirements, and can be the model's processing efficiency and accuracy of data, wherein the training value can be the final value result corresponding to each model indicator.

[0186] In this embodiment, the preset requirements are set in advance.

[0187] In this embodiment, sequence splitting may be splitting the posture features into specific action behaviors corresponding to each moment.

[0188] In this embodiment, the action parameter may be a specific action situation corresponding to each dimension of continuous data at each moment, including facial expressions and hand gestures.

[0189] In this embodiment, the risk probability may be the likelihood that an individual will develop Alzheimer's disease under each dimension of continuous data.

[0190] In this embodiment, the target weight is used to characterize the importance of continuous data of different dimensions when performing risk prediction.

[0191] In this embodiment, the Alzheimer's disease risk prediction value may be the final risk prediction result.

[0192] In this embodiment, the pathological representation may be a specific phenomenon corresponding to the occurrence of Alzheimer's disease in the individual that is finally determined, such as facial condition and limb condition.

[0193] The working principle and beneficial effects of the above technical solution are: by retrieving the corresponding multi-dimensional historical data from the database according to the dimensional information and analyzing the multi-dimensional historical data, it is possible to screen out strongly correlated parameters that are highly correlated with the risk assessment factors from the multi-dimensional historical data, thereby providing reliable training samples for risk prediction; secondly, by determining the correlation attributes between the strongly correlated parameters, it is possible to accurately and effectively train the selected model framework according to the strongly correlated parameters and the corresponding correlation attributes, thereby ensuring the reliability of the final risk prediction model; finally, the individual posture features are sequence-splitting, thereby accurately and effectively locking the individual's behavior path change points, and the obtained behavior path change points are input into the risk prediction model for risk prediction, thereby effectively determining the Alzheimer's disease risk prediction value and pathological characterization, thereby ensuring the accuracy and reliability of Alzheimer's disease risk prediction.

[0194] Example 8:

[0195] Based on Example 7, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the risk prediction unit includes:

[0196] a result acquisition subunit, configured to obtain the association binding result between the obtained Alzheimer's disease risk prediction value and the pathological representation, and at the same time, perform hierarchical matching between the Alzheimer's disease risk prediction value and a preset risk level comparison table, and obtain the Alzheimer's disease risk level based on the hierarchical matching result;

[0197] Record subunit, used to:

[0198] Associating the Alzheimer's disease risk level with the result subordinate tag, and determining the record items and the record representation of each record item based on the subordinate tag;

[0199] A record table is constructed based on the record items and the record representations, and the association binding results of the Alzheimer's disease risk level, the Alzheimer's disease risk prediction value and the pathological representation are recorded and stored in the record table.

[0200] In this embodiment, the preset risk level comparison table is set in advance and is used to record the risk levels corresponding to different prediction values.

[0201] In this embodiment, the subordinate tag may be an association tag for the Alzheimer's disease risk level and the association binding result, thereby facilitating determination of the corresponding relationship between the Alzheimer's disease risk level and the association binding result.

[0202] In this embodiment, the record represents the state or form that the record item needs to present when recording.

[0203] The working principle and beneficial effects of the above technical solution are: by hierarchically matching the Alzheimer's disease risk prediction value with the preset risk level comparison table, the Alzheimer's disease risk level can be accurately and effectively determined; secondly, a record table is constructed to accurately and effectively record the association results of the Alzheimer's disease risk level with the Alzheimer's disease risk prediction value and pathological manifestations in the record table, which provides convenience and basis for risk assessment.

[0204] Example 9:

[0205] Based on Example 1, this example provides an Alzheimer's disease risk prediction system based on continuous data, wherein the prediction module includes:

[0206] A data reading unit is used to read the individual's multi-dimensional continuous data and determine the baseline data value of Alzheimer's disease in each dimension. At the same time, it reads the risk probability value of each dimension of continuous data based on Alzheimer's disease risk prediction;

[0207] a first calculation unit, configured to calculate the accuracy of the result of the Alzheimer's disease risk prediction based on the individual's multi-dimensional continuous data, the baseline data value of the occurrence of Alzheimer's disease in each dimension, and the risk probability value of the Alzheimer's disease risk prediction based on the continuous data in each dimension;

[0208] ;

[0209] in, represents the accuracy of the results of Alzheimer's disease risk prediction; Indicates the dimension sequence value; Indicates the total number of dimensions; Indicates the Continuous data in 3 dimensions; Indicates the The baseline data value of Alzheimer's disease in each dimension; Represents a constant, with a value of 2; Indicates the The influence coefficient corresponding to each dimension is (0.1, 0.2); Indicates the Risk probability value of continuous data based on Alzheimer's disease risk prediction in each dimension;

[0210] The second computing unit is configured to:

[0211] Obtain a total number of predictions for Alzheimer's disease risk prediction, and determine the number of accurate predictions and the number of incorrect predictions among the total number of predictions;

[0212] The prediction efficiency of Alzheimer's disease risk prediction is calculated based on the number of accurate predictions, the number of incorrect predictions, and the accuracy rate of Alzheimer's disease risk prediction results;

[0213] ;

[0214] in, represents the predictive efficiency for Alzheimer's disease risk prediction; Indicates the number of accurate predictions; Indicates the number of wrong predictions; represents the total number of predictions for Alzheimer's disease risk;

[0215] Qualification judgment unit, used for:

[0216] obtaining a prediction efficiency threshold, and comparing the prediction efficiency of the Alzheimer's disease risk prediction with the prediction efficiency threshold to determine whether the prediction of Alzheimer's disease is qualified;

[0217] If the prediction efficiency of the risk prediction of Alzheimer's disease is equal to or greater than the prediction efficiency threshold, then the prediction of Alzheimer's disease is determined to be qualified;

[0218] Otherwise, it is judged to be unqualified for the prediction of Alzheimer's disease;

[0219] The alarm unit is used for performing an alarm operation when it is determined that the prediction of Alzheimer's disease is unqualified.

[0220] In this embodiment, the baseline data value may be pre-set and used to represent the critical value of an indicator of an individual developing Alzheimer's disease.

[0221] In this embodiment, the prediction efficiency threshold may be set in advance and used as a criterion for determining whether the prediction of Alzheimer's disease is qualified.

[0222] In this embodiment, the alarm operation may be one or more of sound, vibration, and light.

[0223] The working principle and beneficial effects of the above technical solution are as follows: by accurately calculating the accuracy rate of the results of Alzheimer's disease risk prediction, it is conducive to using the result accuracy rate as a coefficient of the prediction efficiency of Alzheimer's disease risk prediction, thereby ensuring the accuracy, effectiveness and comprehensiveness of the prediction efficiency calculation of Alzheimer's disease risk prediction; by determining the prediction efficiency threshold, the qualification judgment of Alzheimer's disease prediction is effectively achieved, thereby ensuring the objectivity of the evaluation of Alzheimer's disease prediction results; and by using alarm operation to grasp the prediction status in real time, thereby ensuring the accuracy of Alzheimer's disease prediction.

[0224] Example 10:

[0225] This embodiment provides a method for predicting the risk of Alzheimer's disease based on continuous data. Figure 3 As shown, including:

[0226] Step 1: Collect multi-dimensional continuous data of individuals;

[0227] Step 2: Analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features;

[0228] Step 3: Determine the user's behavior path change points in the target time series based on the posture features, and predict the risk of Alzheimer's disease based on the behavior path change points.

[0229] The working principle and beneficial effects of the above technical solution are as follows: by collecting multi-dimensional continuous data of an individual, reliable data support is provided for Alzheimer's disease risk prediction; secondly, the collected multi-dimensional continuous data is analyzed to accurately and effectively extract the data features of the multi-dimensional continuous data, which facilitates the determination of the individual's posture characteristics; finally, the determined posture characteristics are parsed to accurately determine the change points of the individual's behavior path, thereby effectively determining the individual's action behavior in the target time series, and ultimately achieving accurate and reliable prediction of Alzheimer's disease risk based on the change points of the behavior path, thereby ensuring the accuracy and reliability of Alzheimer's disease risk prediction and greatly facilitating disease treatment.

[0230] Obviously, those skilled in the art may make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if such changes and modifications fall within the scope of the claims and their equivalents, the present invention is intended to include such changes and modifications.

Claims

1. An Alzheimer's disease risk prediction system based on continuous data, characterized in that: include: Data collection module, used to collect multi-dimensional continuous data of individuals; The data analysis module is used to analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features; A prediction module is used to determine the change points of an individual's behavior path in the target time series based on posture features, and to predict the risk of Alzheimer's disease based on the change points of the behavior path; Data analysis modules, including: A dimension identification acquisition unit is used to obtain the dimension identification of the multi-dimensional continuous data, and retrieve the continuous sample data of each dimension in a normal state from a preset database according to the dimension identification of the multi-dimensional continuous data; Target transformation curve construction unit, used for: Reading continuous sample data of each dimension in a normal state, and performing a first mapping on the continuous sample data in a preset rectangular coordinate system according to a time sequence; Obtain target transformation curves of various dimensions according to the first mapping result; A curve analysis unit, configured to analyze the target transformation curve and obtain a reference interval of the target transformation curve; a data feature determination unit, configured to display the continuous data in a rectangular coordinate system of a corresponding dimension, compare the displayed result with a reference interval, and determine the data feature of the multi-dimensional continuous data based on the comparison result; a posture feature determination unit, configured to determine the posture feature of an individual based on the data feature; A posture feature determination unit, comprising: When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the first posture feature is output; When the data feature shows that the continuous data is greater than the reference interval in the rectangular coordinate system of the corresponding dimension, the second posture feature is output; When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the third posture feature is output; Prediction module, including: Data Access Unit, used to: Based on the dimensional information of multi-dimensional continuous data, multi-dimensional historical data is retrieved from the database, and risk assessment factors are determined based on the preset assessment knowledge system of Alzheimer's disease. Extracting data representations of multi-dimensional historical data, and determining the correlation index between the multi-dimensional historical data and risk assessment factors based on the data representations, and determining the multi-dimensional historical data with a correlation index greater than a preset threshold as a strongly correlated parameter; At the same time, based on the preset evaluation knowledge system, the multi-dimensional topological structure between the strongly correlated parameters corresponding to the multi-dimensional historical data is determined, and the correlation attributes between the strongly correlated parameters are obtained based on the multi-dimensional topological structure; Model training unit, used for: Select a model framework from the model library based on the evaluation requirements and extract the model parameters of the model framework; Iteratively train the model parameters based on strongly correlated parameters and correlated attributes, and monitor the training values ​​of each model indicator in real time during the iterative training process. When the training values ​​meet the preset requirements, terminate the iterative training to obtain a risk prediction model. Risk prediction unit, used to: Based on the time development sequence, the posture features corresponding to the multi-dimensional continuous data are separated into sequences to obtain the action parameters of each dimension of the individual continuous data at each moment, and the action parameters of adjacent moments are compared and statistically summarized to obtain the change points of the individual's behavior path under the target time sequence; Based on the risk prediction model, risk analysis is performed on the behavioral path change points to obtain the risk probability of the individual under each dimension of continuous data. Based on the target weight of each dimension of continuous data, the risk probabilities corresponding to multiple dimensions are weighted averaged to obtain the individual's Alzheimer's risk prediction value; At the same time, the individual's Alzheimer's disease pathological representation is determined based on the risk probability and target weight under each dimension of continuous data, and the Alzheimer's disease risk prediction value and the pathological representation are associated and bound.

2. The Alzheimer's disease risk prediction system based on continuous data according to claim 1, characterized in that: Data acquisition module, including: A data collection preparation unit, configured to determine the collection dimensions of individual continuous data based on the collection requirements, and to determine the database corresponding to each collection dimension; The mechanism configuration unit is used to configure data collection mechanisms for different databases and connect the data collection mechanisms with the databases. At the same time, based on the collection requirements, it determines the collection time sequence for continuous data under each collection dimension and configures the parameters of the data collection mechanism based on the collection time sequence; The data collection unit is used to control the data collection mechanism based on the parameter configuration results to collect data from the databases under the corresponding collection dimensions, and to summarize and align the data collection results of each database based on the spatiotemporal characteristics to obtain individual multi-dimensional continuous data.

3. The Alzheimer's disease risk prediction system based on continuous data according to claim 2, characterized in that: Data acquisition unit, including: The data retrieval subunit is used to obtain the obtained multi-dimensional continuous data and visualize the multi-dimensional continuous data in a preset coordinate system in turn; Data preprocessing subunit, used to: Determine the changing trend of continuous data under each collection dimension based on the visual display results, and identify isolated sample points based on the changing trend; Extract the upper and lower sample values ​​of the isolated sample point, and modify the target value of the isolated sample point based on the mean of the upper and lower sample values; Obtaining sample characterization requirements under each collection dimension, and determining a standardized interval for continuous data under each collection dimension based on the sample characterization requirements; The corrected continuous data of each collected dimension are standardized based on the standardized interval.

4. The Alzheimer's disease risk prediction system based on continuous data according to claim 1, characterized in that: Curve analysis unit, including: A second-order derivative operation subunit is used to read the target transformation curve, perform a second-order derivative operation on the target transformation curve, and determine the concave and convex points of the target transformation curve according to the second-order derivative operation result of the target transformation curve; The first baseline determination subunit is configured to: Marking the concave and convex points on the target transformation curve, calculating the first mean of each convex point, and determining the first reference value of the corresponding dimension according to the first mean of each convex point; Performing a second mapping on the first reference value in the target transformation curve to obtain a first reference line; The second baseline determination subunit is configured to: Calculating the second mean of each concave point, and determining the second reference value of the corresponding dimension according to the second mean of each concave point; Performing a third mapping of the second reference value in the target transformation curve to obtain a second reference line; The reference interval determination subunit is configured to determine a reference interval of a target transformation interval according to the first reference line and the second reference line.

5. The Alzheimer's disease risk prediction system based on continuous data according to claim 1, characterized in that: Risk prediction unit, including: a result acquisition subunit, configured to obtain the association binding result between the obtained Alzheimer's disease risk prediction value and the pathological representation, and at the same time, perform hierarchical matching between the Alzheimer's disease risk prediction value and a preset risk level comparison table, and obtain the Alzheimer's disease risk level based on the hierarchical matching result; Record subunit, used to: Associating the Alzheimer's disease risk level with the result subordinate tag, and determining the record items and the record representation of each record item based on the subordinate tag; A record table is constructed based on the record items and the record representations, and the association binding results of the Alzheimer's disease risk level, the Alzheimer's disease risk prediction value and the pathological representation are recorded and stored in the record table.

6. The Alzheimer's disease risk prediction system based on continuous data according to claim 1, characterized in that: Prediction module, including: A data reading unit is used to read the individual's multi-dimensional continuous data and determine the baseline data value of Alzheimer's disease in each dimension. At the same time, it reads the risk probability value of each dimension of continuous data based on Alzheimer's disease risk prediction; a first calculation unit, configured to calculate the accuracy of the result of the Alzheimer's disease risk prediction based on the individual's multi-dimensional continuous data, the baseline data value of the occurrence of Alzheimer's disease in each dimension, and the risk probability value of the Alzheimer's disease risk prediction based on the continuous data in each dimension; ； in, represents the accuracy of the results of Alzheimer's disease risk prediction; Indicates the dimension sequence value; Indicates the total number of dimensions; Indicates the Continuous data in 3 dimensions; Indicates the The baseline data value of Alzheimer's disease in each dimension; Represents a constant, with a value of 2; Indicates the The influence coefficient corresponding to each dimension is (0.1, 0.2); Indicates the Risk probability value of continuous data based on Alzheimer's disease risk prediction in each dimension; The second computing unit is configured to: Obtain a total number of predictions for Alzheimer's disease risk prediction, and determine the number of accurate predictions and the number of incorrect predictions among the total number of predictions; The prediction efficiency of Alzheimer's disease risk prediction is calculated based on the number of accurate predictions, the number of incorrect predictions, and the accuracy rate of Alzheimer's disease risk prediction results; ； in, represents the predictive efficiency for Alzheimer's disease risk prediction; Indicates the number of accurate predictions; Indicates the number of wrong predictions; represents the total number of predictions for Alzheimer's disease risk; Qualification judgment unit, used for: obtaining a prediction efficiency threshold, and comparing the prediction efficiency of the Alzheimer's disease risk prediction with the prediction efficiency threshold to determine whether the prediction of Alzheimer's disease is qualified; If the prediction efficiency of the risk prediction of Alzheimer's disease is equal to or greater than the prediction efficiency threshold, then the prediction of Alzheimer's disease is determined to be qualified; Otherwise, it is judged to be unqualified for the prediction of Alzheimer's disease; The alarm unit is used for performing an alarm operation when it is determined that the prediction of Alzheimer's disease is unqualified.

7. A method for predicting Alzheimer's disease risk based on continuous data, characterized in that: include: Step 1: Collect multi-dimensional continuous data of individuals; Step 2: Analyze the multi-dimensional continuous data, collect the data features of the multi-dimensional continuous data, and output the individual posture features based on the data features; Step 3: Determine the change points of the user's behavior path in the target time series based on the posture features, and predict the risk of Alzheimer's disease based on the change points of the behavior path; Step 1 includes: Obtaining dimension identifiers of multi-dimensional continuous data, and retrieving continuous sample data of each dimension in a normal state from a preset database according to the dimension identifiers of the multi-dimensional continuous data; Reading continuous sample data of each dimension in a normal state, and performing a first mapping on the continuous sample data in a preset rectangular coordinate system according to a time sequence; Obtain target transformation curves of various dimensions according to the first mapping result; Used to analyze the target transformation curve and obtain the reference interval of the target transformation curve; Displaying the continuous data in a rectangular coordinate system of the corresponding dimension, comparing the displayed result with the reference interval, and determining the data characteristics of the multi-dimensional continuous data based on the comparison result; Determine the posture characteristics of the individual based on the data characteristics; When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the first posture feature is output; When the data feature shows that the continuous data is greater than the reference interval in the rectangular coordinate system of the corresponding dimension, the second posture feature is output; When the data feature shows that the continuous data belongs to the reference interval in the rectangular coordinate system of the corresponding dimension, the third posture feature is output; Step 3 includes: Based on the dimensional information of multi-dimensional continuous data, multi-dimensional historical data is retrieved from the database, and risk assessment factors are determined based on the preset assessment knowledge system of Alzheimer's disease. Extracting data representations of multi-dimensional historical data, and determining the correlation index between the multi-dimensional historical data and risk assessment factors based on the data representations, and determining the multi-dimensional historical data with a correlation index greater than a preset threshold as a strongly correlated parameter; At the same time, based on the preset evaluation knowledge system, the multi-dimensional topological structure between the strongly correlated parameters corresponding to the multi-dimensional historical data is determined, and the correlation attributes between the strongly correlated parameters are obtained based on the multi-dimensional topological structure; Select a model framework from the model library based on the evaluation requirements and extract the model parameters of the model framework; Iteratively train the model parameters based on strongly correlated parameters and correlated attributes, and monitor the training values ​​of each model indicator in real time during the iterative training process. When the training values ​​meet the preset requirements, terminate the iterative training to obtain a risk prediction model. Based on the time development sequence, the posture features corresponding to the multi-dimensional continuous data are separated into sequences to obtain the action parameters of each dimension of the individual continuous data at each moment, and the action parameters of adjacent moments are compared and statistically summarized to obtain the change points of the individual's behavior path under the target time sequence; Based on the risk prediction model, risk analysis is performed on the behavioral path change points to obtain the risk probability of the individual under each dimension of continuous data. Based on the target weight of each dimension of continuous data, the risk probabilities corresponding to multiple dimensions are weighted averaged to obtain the individual's Alzheimer's risk prediction value; At the same time, the individual's Alzheimer's disease pathological representation is determined based on the risk probability and target weight under each dimension of continuous data, and the Alzheimer's disease risk prediction value and the pathological representation are associated and bound.

Images