User state evaluation method and system based on heart rate characteristics

Abstract

The invention discloses a user state evaluation method and system based on heart rate characteristics, and relates to the technical field of data analysis, and the method comprises the steps: collecting heart rate time sequence data; performing phase-space reconstruction on the heart rate time series data, and extracting chaotic attractor features representing heart rate complex dynamic characteristics; constructing a heart rate dynamic graph according to the chaotic attractor features, and performing feature aggregation on the dynamic graph by using a graph convolutional network to obtain global topological dynamic features; inputting the global topology dynamic feature, the attention mechanism and the user individual baseline feature into a pre-trained neural network, and outputting an intermediate state value reflecting the current state of the user; and performing adaptive state adjustment on the intermediate state value to generate a final state evaluation result. The method not only improves the precision and robustness of state evaluation, but also provides scientific and reliable technical support for intelligent health management and risk early warning.

Description

Technical Field

[0001] The present invention relates to the technical field of data analysis, and in particular to a user status evaluation method and system based on heart rate characteristics. Background Art

[0002] With the increasing popularity of wearable devices and physiological signal monitoring technologies, heart rate signals, as a key indicator of human health and psychological state, have garnered significant attention in the fields of medicine, health management, and artificial intelligence. Existing health assessment methods primarily rely on indicators such as heart rate and heart rate variability, combined with statistical features or machine learning models, to provide preliminary assessments of a user's physical condition, emotional changes, and health risks. However, as a complex, nonlinear time series signal, the human heart rate signal exhibits a high degree of dynamism and individual variability, driven by multiple factors, including the nervous system, endocrine system, environmental stress, and emotional fluctuations. A single indicator or static statistical model is unlikely to accurately reveal its underlying dynamic mechanisms.

[0003] In recent years, technologies such as dynamical systems theory, chaos theory, time series modeling, and deep learning have been gradually applied to physiological signal analysis, aiming to explore the complex patterns and personalized characteristics in heart rate data. However, in practical applications, how to simultaneously take into account the multi-scale characteristics of the data, individual physiological differences, contextual relevance, and the adaptability of the model remains a challenge faced by academia and industry. Current systems are mostly based on linear analysis, single-scale modeling, or end-to-end deep learning, which makes it difficult to achieve high accuracy, strong interpretability, and continuous self-optimization, limiting the further promotion and implementation of heart rate signals in personalized health monitoring and intelligent intervention scenarios. Therefore, it is urgent to develop new heart rate signal modeling and evaluation methods to meet the needs of efficient, intelligent, and dynamic assessment of user status. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the present invention provides a user status assessment method based on heart rate characteristics to solve the problems of the existing technology in dynamic and accurate assessment and individualized adaptation.

[0006] In order to solve the above technical problems, the present invention provides the following technical solutions: In a first aspect, the present invention provides a user status assessment method based on heart rate characteristics, which includes collecting user heart rate data and performing time series synchronization on the heart rate data to obtain heart rate time series data; Reconstructing the phase space of the heart rate time series data to extract chaotic attractor features that characterize the complex dynamic characteristics of the heart rate; Constructing a heart rate dynamics graph based on the chaotic attractor characteristics, and performing feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features; By obtaining the previous context labels related to the user status and embedding the previous context labels into the attention mechanism; inputting the global topological dynamic features, the attention mechanism and the user's individual baseline features into the pre-trained neural network, and outputting the intermediate state value reflecting the user's current state; Adaptively adjust the state of the intermediate state value to generate a final state evaluation result.

[0007] As a preferred solution of the user status assessment method based on heart rate characteristics of the present invention, the heart rate time series data is represented as: ;in, Indicates in Heart rate value at a time point, Indicates the total number of data points; The phase space reconstruction includes performing feature analysis on the heart rate time series data and generating J groups of data pairs of delay time and embedding dimension number according to the feature analysis results; wherein the jth group of data pairs of delay time and embedding dimension number is expressed as ; represents the delay time of the jth group of data pairs; represents the number of embedding dimensions of the jth group of data pairs; For each data set , using the delayed embedding method, the heart rate data is embedded into dimensional phase space, and obtain the reconstruction vector:

[0008] in, Indicates the jth group of data, reconstructed phase space vector; represents all valid indexes that can be constructed under the j-th set of data pairs; Indicates that under the jth group of data, The data of the first dimension in the reconstructed phase space; Indicates that under the jth group of data, The data of the Nth dimension in the reconstructed phase space; The reconstructed trajectory point set is obtained by integrating the j-th data pair: ; in, Represents the total number of valid trajectory points; integrate the trajectory point sets under all data pairs to obtain the set .

[0009] As a preferred embodiment of the user status assessment method based on heart rate features of the present invention, the feature analysis includes calculating a mutual information function based on the heart rate time series data, and selecting the first minimum value of the mutual information curve as the delay time; after completing the selection of a delay time data, based on the false nearest neighbor method, selecting the minimum embedding dimension when the false nearest neighbor ratio drops to a preset threshold as the embedding dimension; Based on all completed data pairs, the heart rate time series data corresponding to the reconstructed vector is removed from the heart rate time series data, and the mutual information function is recalculated to select the delay time and the embedding dimension using the false nearest neighbor method. When selecting, the data pairs must meet the following constraints: the number is not less than the preset value; After the constraints are met, when the data pairs are generated, the changes in the minimum value of the mutual information are analyzed. If the first minimum value of the mutual information calculated next time is less than the preset value, the next set of data pairs will continue to be generated until the first minimum value of the mutual information calculated next time is not less than the preset value or the preset maximum number of generations is reached.

[0010] As a preferred embodiment of the user state assessment method based on heart rate characteristics of the present invention, the chaotic attractor characteristics include: a dynamic characteristic reflecting the complexity, chaos and nonlinear regulation law of the heart rate system in the trajectory point concentration; and a chaotic characteristic vector in each trajectory point; The chaotic attractor features are extracted for each data pair, and the chaotic attractor features of the jth set of trajectory points after reconstruction are recorded as , including the chaotic feature vector of each trajectory point; the chaotic features under all parameter groups are integrated to form a global multidimensional chaotic feature matrix:

[0011] Among them, each column is a set of features, and each row corresponds to a chaos description dimension.

[0012] As a preferred solution of the user status assessment method based on heart rate characteristics of the present invention, the heart rate dynamics graph includes a trajectory point set for each dimension. , respectively construct the graph structure; each reconstructed phase space trajectory point in the trajectory point set is regarded as a node; the node feature is the corresponding chaotic feature vector; the weight between two nodes is defined as the Euclidean distance between their chaotic feature vectors; By introducing dynamic timing constraints, only adjacent or similar nodes are connected to form J time-series heart rate dynamics graphs; The feature aggregation includes using principal component analysis to identify key nodes of J nodes in the time series heart rate dynamics graph to obtain J groups of key nodes; All key nodes are connected in a new graph structure in time sequence, and the chaotic eigenvectors are recalculated using the key nodes arranged in time sequence. The weight between two nodes is represented by the Euclidean distance between the chaotic eigenvectors.

[0013] As a preferred embodiment of the user state assessment method based on heart rate features of the present invention, the attention mechanism includes: using a pre-trained Bayesian network, inputting the identified user state information and real-time heart rate variability data; and outputting the attention weight distribution result of the user's current state; The intermediate state value is a vector, the direction of which represents the emotion direction, and the value represents the emotion score.

[0014] As a preferred embodiment of the user state assessment method based on heart rate characteristics of the present invention, the adaptive adjustment includes setting the numerical range of the intermediate state value to be in the range of 0-100 as a predicted interval; if the numerical value of the intermediate state value exceeds the predicted interval, intervention is automatically triggered and heart rate data after intervention is collected, and the feedback data is used to optimize the neural network and individual baseline characteristics to achieve closed-loop adaptive optimization of the assessment model; if the numerical value of the intermediate state value exceeds the predicted interval, the state assessment result for each emotion is obtained according to the intermediate state value by using a mapping function preset for each state type, and is used as the current user state; Among them, the initial value of the individual baseline characteristics is the preset standard value; the intervention process includes adjusting the individual baseline characteristics and the weights of the neural network according to the minimum step size, and using the annealing algorithm to optimize the parameters until the value of the intermediate state value exceeds the predicted interval, and the optimization is completed.

[0015] In a second aspect, the present invention provides a user status assessment system based on heart rate characteristics, comprising: a collection unit for collecting heart rate data of a user and performing time series synchronization on the heart rate data to obtain heart rate time series data; a feature extraction unit, which performs phase space reconstruction on the heart rate time series data and extracts chaotic attractor features that characterize the complex dynamic characteristics of the heart rate; A feature aggregation unit constructs a heart rate dynamics graph based on the chaotic attractor features, and performs feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features; The analysis unit obtains a previous context label related to the user state and embeds the previous context label into an attention mechanism; inputs the global topological dynamic features, the attention mechanism, and the user's individual baseline features into a pre-trained neural network, and outputs an intermediate state value reflecting the user's current state; The output unit performs adaptive adjustment on the intermediate state value to generate a final state evaluation result.

[0016] In a third aspect, the present invention provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: when the computer program is executed by the processor, any step of the user status assessment method based on heart rate characteristics as described in the first aspect of the present invention is implemented.

[0017] In a fourth aspect, the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the user status assessment method based on heart rate characteristics as described in the first aspect of the present invention.

[0018] The beneficial effects of the present invention are as follows: by modeling the multi-scale dynamic characteristics of heart rate data and reconstructing the phase space, it is possible to fully explore the complex nonlinear control laws contained in the heart rate signal and improve the ability to identify the user's true state. The introduction of key node identification and topological structure analysis effectively enhances the model's sensitivity to abnormal changes and key physiological events. Combining the attention mechanism with individual baseline characteristics, the adaptive fusion of contextual information and user differences is achieved, making the evaluation results more personalized and interpretable. The present invention can automatically trigger intervention when the state is abnormal, and use intervention feedback to perform closed-loop optimization of the model and baseline parameters, significantly improving the system's self-learning and continuous evolution capabilities. The overall approach not only improves the accuracy and robustness of state assessment, but also provides scientific and reliable technical support for intelligent health management and risk warning. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0020] Figure 1 Flowchart of the user status assessment method based on heart rate features. DETAILED DESCRIPTION

[0021] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0022] In the following description, many specific details are set forth to facilitate a full understanding of the present invention. However, the present invention may also be implemented in other ways different from those described herein. Those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0023] Secondly, the term "one embodiment" or "embodiment" herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in various places throughout this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive of other embodiments.

[0024] Reference Figure 1 , is an embodiment of the present invention, which provides a user status assessment method based on heart rate characteristics, comprising the following steps: S1: Collect the user's heart rate data and perform time series synchronization on the heart rate data to obtain heart rate time series data.

[0025] Furthermore, the heart rate time series data is expressed as: ;in, Indicates in Heart rate value at a time point, Represents the total number of data points.

[0026] By collecting and synchronizing the user's heart rate data, complete, continuous, and standardized heart rate time series data can be obtained, providing basic data support for subsequent dynamic feature extraction, time series modeling, and state assessment. This step helps eliminate problems such as time offset, missing or redundancy during the data collection process, ensuring the consistency and accuracy of the obtained heart rate time series data, thereby laying a solid foundation for systematically analyzing the patterns of heart rate changes and accurately reflecting the user's true physiological state. This design provides a high-quality and reliable data source for subsequent complex feature analysis, modeling, and assessment, improving the scientific nature and effectiveness of the entire assessment process.

[0027] S2: Reconstructing the phase space of the heart rate time series data to extract chaotic attractor features that characterize the complex dynamic characteristics of the heart rate.

[0028] The phase space reconstruction includes performing feature analysis on the heart rate time series data and generating J groups of data pairs of delay time and embedding dimension number according to the feature analysis results; wherein the jth group of data pairs of delay time and embedding dimension number is expressed as ; represents the delay time of the jth group of data pairs; Represents the number of embedding dimensions of the jth group of data pairs.

[0029] For each data set , using the delayed embedding method, the heart rate data is embedded into dimensional phase space, and obtain the reconstruction vector:

[0030] in, Indicates the jth group of data, reconstructed phase space vector; Represents all valid indexes that can be constructed under the j-th set of data pairs. Indicates that under the jth group of data, The data of the first dimension in the reconstructed phase space; Indicates that under the jth group of data, The data of the Nth dimension in the reconstructed phase space. The reconstructed trajectory point set is obtained by integrating the jth set of data pairs: .in, Represents the total number of valid trajectory points; integrate the trajectory point sets under all data pairs to obtain the set .

[0031] Furthermore, based on the heart rate time series data, the mutual information function is calculated, and the first minimum value of the mutual information curve is selected as the delay time; after completing the selection of a delay time data, based on the false nearest neighbor method, the minimum embedding dimension when the false nearest neighbor ratio drops to a preset threshold is selected as the embedding dimension.

[0032] According to all completed data pairs, the heart rate time series data corresponding to the reconstructed vector is eliminated from the heart rate time series data, and the mutual information function is recalculated in the eliminated heart rate time series data to select the delay time and the false nearest neighbor method is used to select the embedding dimension.

[0033] During selection, the data pairs must meet the constraint that their number must be no less than a preset value. Once these constraints are met, the minimum mutual information value is analyzed as the data pairs are generated. If the first minimum mutual information value calculated in the next iteration is less than the preset value, the next set of data pairs is generated until the first minimum mutual information value calculated in the next iteration is no less than the preset value or the preset maximum number of generation attempts is reached.

[0034] It should be noted that by reconstructing the phase space of heart rate time series data using multiple sets of parameters, the complex dynamic behavior of heart rate signals at different scales can be fully captured. Using the mutual information function to select the optimal delay time and combining it with the false nearest neighbor method to determine the embedding dimension helps ensure the accuracy and scientific nature of the reconstructed phase space and avoids feature omissions or redundancy. The progressive generation of data pairs with multiple sets of parameters not only enhances the adaptability to the multi-scale and nonlinear structure of heart rate signals, but also improves the richness and discriminability of dynamic features. By eliminating used data after each round of reconstruction, information overlap can be reduced and the independence and representativeness of features can be improved. Dynamically adjusting the data pair generation process can prevent invalid parameter expansion or redundant feature dimensions, ensuring that the final feature set is both comprehensive and efficient.

[0035] The chaotic attractor features include dynamical characteristics that reflect the complexity, chaos, and nonlinear regulation of the heart rate system at the trajectory point concentration; these characteristics are reflected as chaotic eigenvectors at each trajectory point. In this embodiment, the chaotic attractor features refer to a set of dynamical features that reflect the complexity, chaos, and nonlinear regulation of the heart rate system, extracted from the chaotic attractor trajectory formed in the high-dimensional phase space after the phase space reconstruction of the heart rate time series data. These features include, but are not limited to, correlation dimension, maximum Lyapunov exponent, approximate entropy, sample entropy, average radius of the attractor trajectory, traversal length, spatial density distribution, and topological parameters of the attractor network. This set of features serves as a key input for assessing and identifying a user's heart rate status, and can comprehensively and quantitatively reveal the nonlinear dynamic mechanisms behind heart rate changes. The correlation dimension quantifies the fractal complexity of the attractor trajectory, reflecting the dynamic degrees of freedom of the heart rate system. The maximum Lyapunov exponent measures the attractor's sensitivity to initial perturbations and reveals the chaotic characteristics of heart rate regulation. A positive value indicates chaotic behavior in the system. Approximate entropy and sample entropy are used to characterize the complexity and unpredictability of attractors. Larger values ​​indicate more disordered and unpredictable heart rate sequences. Attractor trajectory statistics, including the attractor's average radius, traversal length, and density distribution in phase space, reflect the spatial distribution structure and activity level of the trajectory.

[0036] The attractor network topological characteristics include node distribution, clustering coefficient, average path length and other indicators of the network constructed based on the attractor trajectory, which are used to reveal the multi-scale correlation and regulation mechanism of the heart rate dynamics system.

[0037] The chaotic attractor features are extracted for each data pair, and the chaotic attractor features of the jth set of trajectory points after reconstruction are recorded as , including the chaotic feature vector of each trajectory point; the chaotic features under all parameter groups are integrated to form a global multidimensional chaotic feature matrix:

[0038] Among them, each column is a set of features, and each row corresponds to a chaotic description dimension. By extracting the chaotic attractor features of the reconstructed trajectory point set under each set of parameters and integrating them into a global multi-dimensional chaotic feature matrix, comprehensive and rich data support can be provided for subsequent key node screening and time series dynamic graph structure reconstruction. As a comprehensive expression carrier of multiple sets of parameters and multi-scale dynamic characteristics, this matrix enables subsequent layer-by-layer analysis steps such as principal component analysis and key node identification to systematically discover core areas and abnormal change points at different dynamic scales based on multi-dimensional information, thereby improving the scientificity and effectiveness of feature screening and structural optimization. This design lays a solid foundation for dynamic characteristics for more accurate state assessment and model adaptive optimization.

[0039] S3: constructing a heart rate dynamics graph based on the chaotic attractor characteristics, and performing feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features.

[0040] The heart rate dynamics graph includes a trajectory point set for each dimension , respectively construct the graph structure; each reconstructed phase space trajectory point in the trajectory point set is regarded as a node; the node feature is the corresponding chaotic feature vector; the weight between two nodes is defined as the Euclidean distance between their chaotic feature vectors.

[0041] By introducing dynamic timing constraints, only adjacent or similar nodes are connected to form J time-series heart rate dynamics graphs. “Only connecting adjacent or similar nodes” means that when constructing the heart rate dynamics graph, only connections are established between reconstructed trajectory points that are continuous in time series or whose time interval is less than a preset threshold. Specifically, assuming that the time series indexes of the trajectory points are and ,like (in is the set time proximity threshold, usually 1 or less than a small constant), then an edge is established for these two nodes in the dynamic graph.

[0042] The feature aggregation method involves using principal component analysis to identify key nodes for each of J time-series heart rate dynamics graphs, thereby obtaining J groups of key nodes. In this embodiment, each key node is associated with a non-key node in its respective time-series heart rate dynamics graph that is directly connected and has a weight greater than a preset value. When constructing a new graph structure, the relationship between each two connected nodes is analyzed. If, after adding a "node associated with either node," both nodes can establish a high-weighted relationship with the newly added node, the newly added node is retained. This allows for node replenishment.

[0043] All key nodes are connected in a new graph structure in time sequence, and the chaotic eigenvectors are recalculated using the key nodes arranged in time sequence. The weight between two nodes is represented by the Euclidean distance between the chaotic eigenvectors.

[0044] The point is that by constructing a heart rate dynamics graph based on chaotic attractor characteristics and introducing dynamic timing constraints, only connecting nodes with similar timing, we can accurately and precisely reflect the continuous evolution of heart rate signals under multi-scale dynamics, avoiding redundant or meaningless connections between unrelated nodes. Each node uses a chaotic eigenvector as an attribute, and Euclidean distance as an edge weight, ensuring the graph structure's expressiveness for complex dynamic patterns.

[0045] Principal component analysis (PCA) is used to screen key nodes, automatically identifying the regions most representative of dynamic evolution from multi-dimensional information, effectively highlighting outliers or regulatory cores. Furthermore, by supplementing non-key nodes that are strongly correlated with key nodes, the integrity of the local topology and global structure can be corrected, avoiding information omissions caused by simple screening.

[0046] Ultimately, reconnecting key nodes and calculating their features in a time-sequential manner not only improves the new graph structure's ability to depict the main dynamics, but also provides a more targeted and discriminative high-dimensional feature foundation for subsequent feature aggregation and global state assessment. This design ensures that the assessment model can both grasp global evolutionary trends and keenly perceive local anomalies, achieving the goal of highly accurate and highly interpretable personalized dynamic assessment.

[0047] S4: By obtaining the previous context labels related to the user status and embedding the previous context labels into the attention mechanism; the global topological dynamic features, the attention mechanism and the user's individual baseline features are input into the pre-trained neural network, and the intermediate state value reflecting the user's current state is output.

[0048] The attention mechanism includes using a pre-trained Bayesian network to input recognized user state information (the recognition result of the user state before the current moment) and real-time heart rate variability data (here, a data sequence of a shorter time interval, representing the data sequence of the current and previous period); and outputting the attention weight distribution result for the user's current state.

[0049] By introducing previous contextual labels related to the user's state and embedding them into the attention mechanism, the model effectively captures the dynamic connection between historical state changes and current physiological signals, thereby enhancing continuous perception of user behavior and psychological trends. Using a pre-trained Bayesian network, historical state information is jointly modeled with real-time heart rate variability data, outputting the attention weights of different features on the current state, enabling automatic allocation of feature contributions and causal inference.

[0050] By integrating global topological dynamic features, attention weights, and individual user baseline characteristics into the neural network, we achieve a deep fusion of multimodal and multi-scale information, improving the precise identification and individualized adaptation of the user's current state. Outputting intermediate state values ​​in vector form enables the model to not only provide directional judgments on emotion types but also quantitatively score the degree of emotion or health risk, facilitating subsequent decision-making and dynamic intervention. This design ensures that the model's assessment of user status is both comprehensive and detailed, combining continuity, personalization, and interpretability, laying a solid foundation for efficient and intelligent health management.

[0051] The intermediate state value is a vector, the direction of which represents the emotion direction, and the value represents the emotion score.

[0052] S5: Adaptively adjust the intermediate state value to generate a final state evaluation result.

[0053] Furthermore, the adaptive adjustment includes setting the numerical range of the intermediate state value to 0-100 as the interval that meets the prediction; if the numerical value of the intermediate state value exceeds the interval that meets the prediction, the intervention is automatically triggered and the heart rate data after the intervention is collected, and the feedback data is used to optimize the neural network and individual baseline characteristics to achieve closed-loop adaptive optimization of the evaluation model; if the numerical value of the intermediate state value exceeds the interval that meets the prediction, then according to the intermediate state value, through the preset mapping function for each state type (the preset conversion function converts the vector into the evaluation result for each emotion, and at the same time reflects the "component" of each emotion), the state evaluation result under each emotion is obtained as the current user state.

[0054] Among them, the initial value of the individual baseline characteristics is the preset standard value; the intervention process includes adjusting the individual baseline characteristics and the weights of the neural network according to the minimum step size, and using the annealing algorithm to optimize the parameters until the value of the intermediate state value exceeds the predicted interval, and the optimization is completed.

[0055] Parameter fine-tuning, combined with an annealing algorithm, strikes a balance between convergence efficiency and global optimization, enabling the model to quickly adapt to recent changes in user status and achieving closed-loop adaptive optimization. A pre-set mapping function transforms the intermediate state vector into a multi-component, fine-grained quantitative result for various emotions or health risks, providing a more actionable basis for clinical intervention, health reminders, and more. This not only improves the system's sensitivity and adaptability to abnormal conditions, but also ensures the individualized evolution and continuous optimization of evaluation results over the long term, ultimately achieving the goal of efficient, intelligent, and reliable user status management.

[0056] This embodiment further provides a user status assessment system based on heart rate characteristics, including: The collecting unit collects the user's heart rate data and performs time series synchronization on the heart rate data to obtain heart rate time series data.

[0057] The feature extraction unit reconstructs the phase space of the heart rate time series data and extracts the chaotic attractor features that characterize the complex dynamic characteristics of the heart rate.

[0058] A feature aggregation unit constructs a heart rate dynamics graph based on the chaotic attractor features, and performs feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features.

[0059] The analysis unit obtains the previous context labels related to the user status and embeds the previous context labels into the attention mechanism; inputs the global topological dynamic features, the attention mechanism and the user's individual baseline features into the pre-trained neural network, and outputs an intermediate state value reflecting the user's current state.

[0060] The output unit performs adaptive adjustment on the intermediate state value to generate a final state evaluation result.

[0061] This embodiment also provides a computer device, which is suitable for the user status assessment method based on heart rate characteristics, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute computer-executable instructions to implement the user status assessment method based on heart rate characteristics proposed in the above embodiment.

[0062] The computer device may be a terminal, comprising a processor, memory, a communication interface, a display, and an input device connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores an operating system and computer programs. The internal memory provides an environment for the operating system and computer programs stored in the non-volatile storage media. The communication interface of the computer device is used to communicate with external terminals via wired or wireless communication. Wireless communication may be achieved via Wi-Fi, a carrier network, NFC (near-field communication), or other technologies. The display of the computer device may be a liquid crystal display or an electronic ink display. The input device may be a touchscreen overlay on the display, buttons, a trackball, or a touchpad on the computer device housing, or an external keyboard, touchpad, or mouse.

[0063] This embodiment also provides a storage medium having a computer program stored thereon. When the program is executed by a processor, the method for evaluating user status based on heart rate characteristics as proposed in the above embodiment is implemented. The storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0064] In summary, the present invention integrates multiple cutting-edge technologies, including complex dynamics modeling, graph neural networks, attention mechanisms, and adaptive optimization. This method first comprehensively explores the chaotic attractor characteristics of heart rate time series data at different scales through multi-parameter phase space reconstruction, and uses this to construct a multidimensional heart rate dynamics graph. Principal component analysis is used to automatically identify key nodes, and combined with dynamic time series constraints, it accurately characterizes the continuity and local anomalies of heart rate changes. Using a pre-trained Bayesian network, historical state and real-time heart rate variability data are fused, and attention weights for different features are output to achieve causal inference and information allocation.

[0065] The model integrates global topological dynamic features, attention mechanisms, and individual baseline characteristics into a neural network, outputting intermediate state values ​​that reflect the direction and score of sentiment. In the event of abnormal states, the system automatically triggers intelligent intervention and implements closed-loop adaptive optimization of neural network weights and individual baseline parameters through incremental feedback and annealing algorithms.

[0066] It significantly improves the accuracy, personalization and dynamic response capabilities of user status assessment. It can not only reflect complex physiological and psychological changes in detail, but also continuously optimize, and has good application prospects and promotion value.

[0067] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention, which should all be included in the scope of the claims of the present invention.

Claims

1. A user status assessment method based on heart rate characteristics, characterized by: include, Collecting the user's heart rate data and performing time series synchronization on the heart rate data to obtain heart rate time series data; Reconstructing the phase space of the heart rate time series data to extract chaotic attractor features that characterize the complex dynamic characteristics of the heart rate; Constructing a heart rate dynamics graph based on the chaotic attractor characteristics, and performing feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features; By obtaining the previous context labels related to the user status and embedding the previous context labels into the attention mechanism; inputting the global topological dynamic features, the attention mechanism and the user's individual baseline features into the pre-trained neural network, and outputting the intermediate state value reflecting the user's current state; Adaptively adjust the state of the intermediate state value to generate a final state evaluation result.

2. The user status assessment method based on heart rate characteristics according to claim 1, wherein: The heart rate time series data is expressed as: ;in, Indicates in Heart rate value at a time point, Indicates the total number of data points; The phase space reconstruction includes performing feature analysis on the heart rate time series data and generating J groups of data pairs of delay time and embedding dimension number according to the feature analysis results; wherein the jth group of data pairs of delay time and embedding dimension number is expressed as ; represents the delay time of the jth group of data pairs; represents the number of embedding dimensions of the jth group of data pairs; For each data set , using the delayed embedding method, the heart rate data is embedded into dimensional phase space, and obtain the reconstruction vector: in, Indicates the jth group of data, reconstructed phase space vector; represents all valid indexes that can be constructed under the j-th set of data pairs; Indicates that under the jth group of data, The data of the first dimension in the reconstructed phase space; Indicates that under the jth group of data, The data of the Nth dimension in the reconstructed phase space; The reconstructed trajectory point set is obtained by integrating the j-th data pair: ; in, Represents the total number of valid trajectory points; integrate the trajectory point sets under all data pairs to obtain the set .

3. The user status assessment method based on heart rate characteristics according to claim 2, characterized in that: The feature analysis includes calculating a mutual information function based on the heart rate time series data, and selecting a first minimum value of the mutual information curve as the delay time; After selecting a delay time data, based on the false nearest neighbor method, the minimum embedding dimension when the false nearest neighbor ratio drops to a preset threshold is selected as the embedding dimension; Based on all completed data pairs, the heart rate time series data corresponding to the reconstructed vector is removed from the heart rate time series data, and the mutual information function is recalculated to select the delay time and the embedding dimension using the false nearest neighbor method. When selecting, the data pairs must meet the following constraints: the number is not less than the preset value; After the constraints are met, when the data pairs are generated, the changes in the minimum value of the mutual information are analyzed. If the first minimum value of the mutual information calculated next time is less than the preset value, the next set of data pairs will continue to be generated until the first minimum value of the mutual information calculated next time is not less than the preset value or the preset maximum number of generations is reached.

4. The user status assessment method based on heart rate characteristics according to claim 3, wherein: The chaotic attractor characteristics include the concentration of trajectory points, which reflects the dynamic characteristics of the complexity, chaos and nonlinear regulation of the heart rate system; In each trajectory point, it is reflected as a chaotic eigenvector; The chaotic attractor features are extracted for each data pair, and the chaotic attractor features of the jth set of trajectory points after reconstruction are recorded as , including the chaotic feature vector of each trajectory point; the chaotic features under all parameter groups are integrated to form a global multidimensional chaotic feature matrix: Among them, each column is a set of features, and each row corresponds to a chaos description dimension.

5. The user status assessment method based on heart rate characteristics according to claim 4, characterized in that: The heart rate dynamics graph includes a trajectory point set for each dimension , respectively construct the graph structure; each reconstructed phase space trajectory point in the trajectory point set is regarded as a node; the node feature is the corresponding chaotic feature vector; the weight between two nodes is defined as the Euclidean distance between their chaotic feature vectors; By introducing dynamic timing constraints, only adjacent or similar nodes are connected to form J time-series heart rate dynamics graphs; The feature aggregation includes using principal component analysis to identify key nodes of J nodes in the time series heart rate dynamics graph to obtain J groups of key nodes; All key nodes are connected in a new graph structure in time sequence, and the chaotic eigenvectors are recalculated using the key nodes arranged in time sequence. The weight between two nodes is represented by the Euclidean distance between the chaotic eigenvectors.

6. The user status assessment method based on heart rate characteristics according to claim 5, characterized in that: The attention mechanism includes using a pre-trained Bayesian network, inputting the recognized user status information and real-time heart rate variability data, and outputting the attention weight distribution result of the user's current status; The intermediate state value is a vector, the direction of which represents the emotion direction, and the value represents the emotion score.

7. The user status assessment method based on heart rate characteristics according to claim 6, characterized in that: The adaptive adjustment includes setting the numerical range of the intermediate state value to be 0-100 as the interval that meets the prediction; If the value of the intermediate state value exceeds the predicted range, intervention is automatically triggered and heart rate data after intervention is collected. The feedback data is used to optimize the neural network and individual baseline characteristics, achieving closed-loop adaptive optimization of the evaluation model; If the value of the intermediate state value exceeds the predicted interval, then according to the intermediate state value, the state evaluation result of each emotion is obtained by using the mapping function preset for each state type, and is used as the current user state; Among them, the initial values ​​of individual baseline characteristics are preset standard values; The intervention process includes adjusting the individual baseline characteristics and the weights of the neural network according to the minimum step size, and optimizing the parameters using the annealing algorithm until the value of the intermediate state exceeds the predicted range, completing the optimization.

8. A user status assessment system based on heart rate characteristics, based on the user status assessment method based on heart rate characteristics according to any one of claims 1 to 7, characterized in that: include, A collection unit collects the user's heart rate data and performs time series synchronization on the heart rate data to obtain heart rate time series data; a feature extraction unit, which performs phase space reconstruction on the heart rate time series data and extracts chaotic attractor features that characterize the complex dynamic characteristics of the heart rate; A feature aggregation unit constructs a heart rate dynamics graph based on the chaotic attractor features, and performs feature aggregation on the dynamics graph using a graph convolutional network to obtain global topological dynamic features; The analysis unit obtains a previous context label related to the user state and embeds the previous context label into an attention mechanism; inputs the global topological dynamic features, the attention mechanism, and the user's individual baseline features into a pre-trained neural network, and outputs an intermediate state value reflecting the user's current state; The output unit performs adaptive adjustment on the intermediate state value to generate a final state evaluation result.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the user status assessment method based on heart rate characteristics according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the user status assessment method based on heart rate characteristics according to any one of claims 1 to 7 are implemented.

Images