Sleep quality monitoring management system and method based on neural network

Abstract

The invention provides a sleep quality monitoring management system and method based on a neural network, relates to the field of intelligent monitoring, and uses pulse monitoring equipment to collect pulse data of a monitored object in sleep in real time, form a time queue and transmit the time queue to an intelligent terminal. The intelligent terminal firstly screens out abnormal data to ensure data accuracy; and then classifying the screened pulse data, establishing a mapping relation between a classification result and a sleep stage, and generating a sleep stage mapping result. And finally, evaluating the sleep quality of the monitored object according to the mapping results, and providing targeted sleep improvement suggestions. In this way, through accurate data processing and analysis, the accuracy of sleep monitoring and the quality of personalized services can be effectively improved.

Description

Technical Field

[0001] The present application relates to the field of intelligent monitoring, and more particularly to a sleep quality monitoring management system and method based on a neural network in an embodiment of the present application. Background Art

[0002] Sleep quality directly affects a person's health and quality of life. With the accelerated pace of modern life, more and more people are facing sleep problems, such as insomnia, sleep apnea syndrome, etc. Therefore, a simple and efficient sleep quality monitoring method has become particularly important. Patent CN103654744A discloses a sleep quality monitoring method and system, which obtains the pulse data of the monitored person during sleep through a pulse monitoring device, and transmits the data to an intelligent terminal. The intelligent terminal uses a clustering function algorithm to classify the pulse data, and establishes a mapping relationship between the classification results and the different stages of sleep. Based on this mapping relationship, the intelligent terminal can judge the sleep quality of the monitored person, and extract corresponding suggestions from the sleep management database to provide to the user.

[0003] In actual applications, pulse data may have abnormal values ​​due to various external or internal factors, such as poor sensor contact, sudden user movements, or environmental interference. If these abnormal data are not processed and used directly for analysis, they may lead to misjudgment of sleep stages, thus affecting the accuracy of the final sleep quality assessment.

[0004] This patent uses a fuzzy algorithm to filter out abnormal data in pulse data. However, fuzzy algorithms rely on predefined rules or models when dealing with uncertainty problems, which limits their adaptability to unforeseen situations, especially in complex and changeable practical application scenarios. In addition, since parameter adjustment requires manual intervention, it is difficult for the system to automatically optimize to adapt to individual differences or environmental changes, resulting in limited flexibility and response speed. In addition, the static characteristics of the fuzzy logic system make it impossible to effectively learn emerging patterns or trends, reducing the accuracy and reliability of long-term use. This limitation is particularly evident when facing personalized needs and dynamic environments, making it difficult to provide continuous and accurate services.

[0005] Therefore, an optimized sleep quality monitoring and management solution is desired. Summary of the invention

[0006] In order to solve the above technical problems, this application is proposed.

[0007] According to one aspect of the present application, a sleep quality monitoring and management method based on a neural network is provided, which includes: using a pulse monitoring device to collect pulse data of a monitored object in real time during sleep to obtain a time queue of pulse data, and transmitting the time queue of pulse data to an intelligent terminal; in the intelligent terminal, screening the time queue of pulse data for abnormal data to obtain a time queue of screened pulse data; in the intelligent terminal, classifying the time queue of screened pulse data to obtain a classification result, and establishing a mapping relationship between the classification result and the sleep stage to obtain a sleep stage mapping result; in the intelligent terminal, obtaining the sleep quality of the monitored object according to the sleep stage mapping result, and providing sleep improvement suggestions for the monitored object according to the sleep quality; wherein, the pulse The method comprises: extracting the initial screening abnormal pulse data from the time queue of the pulse data, and dividing the time queue of the pulse data to obtain the time queue of the forward pulse data and the time queue of the backward pulse data; performing semantic association temporal coding on the time queue of the forward pulse data and the time queue of the backward pulse data according to the initial screening abnormal pulse data to obtain the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features; performing feature interaction of the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features with a dynamic memory mechanism to obtain the pulse data semantic association forward and backward interactive fusion features, and obtaining the recognition result according to the pulse data semantic association forward and backward interactive fusion features.

[0008] According to another aspect of the present application, a sleep quality monitoring and management system based on a neural network is provided, which includes: a pulse data time queue extraction module, which is used to extract initial screening abnormal pulse data from the time queue of the pulse data, and divide the time queue of the pulse data to obtain a time queue of forward pulse data and a time queue of backward pulse data; a semantic association temporal encoding module, which is used to perform semantic association temporal encoding on the time queue of the forward pulse data and the time queue of the backward pulse data respectively according to the initial screening abnormal pulse data to obtain pulse data semantic association forward temporal features and pulse data semantic association backward temporal features; a dynamic memory mechanism feature interaction module, which is used to perform feature interaction of the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features with a dynamic memory mechanism to obtain pulse data semantic association forward and backward interactive fusion features, and obtain recognition results based on the pulse data semantic association forward and backward interactive fusion features.

[0009] Compared with the prior art, the present application provides a sleep quality monitoring management system and method based on a neural network, which extracts and divides the initial screening abnormal pulse data from the time queue of the pulse data, and then performs semantic association temporal coding, so as to intelligently judge whether to determine the initial screening abnormal pulse data as abnormal pulse data based on the dynamic memory interaction representation between the forward temporal features of the semantic association of the pulse data and the backward temporal features of the semantic association of the pulse data. In this way, it can automatically adapt to individual physiological differences and environmental changes, reduce dependence on manual intervention to adjust parameters, and enhance the robustness of the system when encountering unknown situations or new patterns. At the same time, it can continuously learn new patterns and trends from new data to improve performance in meeting personalized needs and adapting to dynamic environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] By describing the embodiments of the present application in more detail in conjunction with the accompanying drawings, the above and other purposes, features and advantages of the present application will become more apparent. The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the embodiments of the present application, they are used to explain the present application and do not constitute a limitation of the present application. In the accompanying drawings, the same reference numerals generally represent the same components or steps.

[0011] Figure 1 Flow chart of a sleep quality monitoring and management method based on a neural network according to an embodiment of the present application.

[0012] Figure 2 The present invention is a flowchart of filtering abnormal data from the time queue of the pulse data to obtain the filtered pulse data in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application.

[0013] Figure 3 The present invention is a data flow diagram of filtering out abnormal data in the time queue of the pulse data to obtain the pulse data after filtering out in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application.

[0014] Figure 4 This is a flowchart of performing semantic association temporal encoding on the time queue of the forward pulse data and the time queue of the backward pulse data according to the initial screening abnormal pulse data in the neural network-based sleep quality monitoring and management method according to an embodiment of the present application to obtain semantic association forward temporal features of the pulse data and semantic association backward temporal features of the pulse data.

[0015] Figure 5A flowchart is provided for performing feature interaction of the pulse data semantically associated forward time series features and the pulse data semantically associated backward time series features according to an embodiment of the present application in the neural network-based sleep quality monitoring and management method to obtain the pulse data semantically associated forward and backward interactive fusion features, and obtaining the recognition result based on the pulse data semantically associated forward and backward interactive fusion features.

[0016] Figure 6 The present invention is a flowchart of performing feature interaction of the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector using a dynamic memory mechanism in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application to obtain a pulse data semantically associated forward and backward interactive fusion feature vector.

[0017] Figure 7 4 is a system block diagram of a sleep quality monitoring and management system based on a neural network according to an embodiment of the present application. DETAILED DESCRIPTION

[0018] Various exemplary embodiments, features and aspects of the present application will be described in detail below with reference to the accompanying drawings. The same reference numerals in the accompanying drawings represent elements with the same or similar functions. Although various aspects of the embodiments are shown in the accompanying drawings, the drawings are not necessarily drawn to scale unless otherwise specified.

[0019] Sleep quality is crucial to health and quality of life, especially in the context of the accelerated pace of modern life, many people face sleep problems. In order to monitor sleep quality, patent CN103654744A proposes a method for analyzing sleep quality through pulse data. The pulse monitoring device collects data and transmits it to the intelligent terminal, which classifies the data through a clustering algorithm, establishes a corresponding relationship with different stages of sleep, and thus evaluates sleep quality and gives improvement suggestions. However, pulse data may be subject to various interferences, such as poor sensor contact or user activity, which will cause abnormal data. If these abnormal data are not processed, it may affect the judgment of sleep stages and the accuracy of sleep quality assessment. The patent uses a fuzzy algorithm to filter abnormal data, but the fuzzy algorithm relies on preset rules, lacks flexibility, and cannot adapt to new situations. Especially in a dynamically changing environment, the fuzzy algorithm requires manual adjustment and cannot learn new patterns. Long-term use may affect accuracy and reliability, and it is difficult to provide continuous and accurate sleep quality monitoring.

[0020] In response to the above technical problems, this application proposes a sleep quality monitoring and management method based on neural network. Figure 1 FIG. 1 is a flowchart of a sleep quality monitoring and management method based on a neural network according to an embodiment of the present application. Figure 1As shown, according to the neural network-based sleep quality monitoring and management method of the embodiment of the present application, the method includes: S110, using a pulse monitoring device to collect pulse data of the monitored object during sleep in real time to obtain a time queue of pulse data, and transmitting the time queue of pulse data to a smart terminal; S120, at the smart terminal, screening the time queue of pulse data for abnormal data to obtain a time queue of screened pulse data; S130, at the smart terminal, classifying the time queue of screened pulse data to obtain a classification result, and establishing a mapping relationship between the classification result and the sleep stage to obtain a sleep stage mapping result; S140, at the smart terminal, obtaining the sleep quality of the monitored object according to the sleep stage mapping result, and providing sleep improvement suggestions to the monitored object based on the sleep quality.

[0021] In the above sleep quality monitoring and management method based on neural network, the step S110 uses a pulse monitoring device to collect the pulse data of the monitored object in real time during sleep to obtain a time queue of the pulse data, and transmits the time queue of the pulse data to the intelligent terminal. It should be understood that the pulse monitoring device usually includes a sensor module, a data acquisition processing unit and a communication module. During sleep, the device senses the pulse signal in real time by contacting the body of the monitored object (for example, through a sensor worn on the wrist, fingertips, earlobe or chest). Among them, the acquisition of the pulse signal is achieved by a sensor (such as a photoelectric volume pulse wave sensor, an infrared sensor, or a capacitive sensor, etc.), which can accurately detect the changes in the pulse wave and the pulsation frequency of blood flow. In this process, the pulse monitoring device will continuously record the pulse wave generated by each heartbeat, usually sampling at a high frequency (such as multiple sampling points per second), thereby forming a time queue of pulse data. This time queue represents the pulse changes of the monitored object during sleep, including multiple parameters such as pulse frequency, pulse interval, waveform changes, etc. These data can reflect the physiological state of the monitored person during sleep, such as whether there is irregular heart rate, rapid heartbeat or too slow pulse, and thus provide an in-depth analysis of the health status.

[0022] Once the pulse data is effectively collected and processed, the next task is to transmit the data to the smart terminal for further analysis and display. At this time, the pulse monitoring device needs to have wireless communication capabilities, and common methods include short-range wireless transmission technologies such as Bluetooth, Wi-Fi or Zigbee. Through these wireless communication technologies, the monitoring device can send the time queue of pulse data to the smart terminal in real time. After the smart terminal receives this data, it can display, store and analyze it through the corresponding health management application.

[0023] In the above-mentioned sleep quality monitoring and management method based on neural network, in step S120, in the intelligent terminal, the time queue of the pulse data is screened for abnormal data to obtain the time queue of the pulse data after screening. It should be understood that, considering that the pulse data is usually recorded in the form of a time series, there may be abnormal values ​​in these data due to various external factors (such as measurement errors, equipment failures, interference, etc.). These abnormal values ​​may appear as sudden changes in pulse frequency, unreasonable fluctuations, or values ​​that are significantly deviated from the normal physiological range. If they are not screened out, they will affect the authenticity and validity of the data, and even lead to incorrect health judgments. The screened data set can provide a more realistic basis for subsequent health analysis and help users better monitor their physical condition.

[0024] Accordingly, in the process of filtering out abnormal data from the time queue of the pulse data to obtain the pulse data after filtering out, the technical concept of the present application is to extract the initial screening abnormal pulse data from the time queue of the pulse data by using a data analysis and processing algorithm based on a neural network, and to divide the time queue of the pulse data, and then to perform semantic association temporal coding on the time queue of the forward pulse data and the time queue of the backward pulse data after the division, so as to intelligently judge whether to determine the initial screening abnormal pulse data as abnormal pulse data according to the dynamic memory interaction representation between the forward temporal features of the semantic association of the pulse data and the backward temporal features of the semantic association of the pulse data. In this way, it can automatically adapt to individual physiological differences and environmental changes, reduce the dependence on manual intervention to adjust parameters, and enhance the robustness of the system when encountering unknown situations or new patterns. At the same time, it has a strong dynamic learning ability, and can continuously learn new patterns and trends from new data, thereby improving its performance in meeting personalized needs and adapting to dynamic environments.

[0025] Figure 2 The present invention is a flowchart of filtering abnormal data from the time queue of the pulse data to obtain the filtered pulse data in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application. Figure 3 The figure is a data flow diagram of filtering abnormal data from the pulse data time queue in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application to obtain the pulse data after filtering. Figure 2 and Figure 3As shown, in the step S120, the time queue of the pulse data is screened for abnormal data to obtain the screened pulse data, including: S121, extracting the initial screening abnormal pulse data from the time queue of the pulse data, and dividing the time queue of the pulse data to obtain the time queue of forward pulse data and the time queue of backward pulse data; S122, according to the initial screening abnormal pulse data, respectively performing semantic association temporal encoding on the time queue of the forward pulse data and the time queue of the backward pulse data to obtain the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features; S123, performing feature interaction of the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features with a dynamic memory mechanism to obtain the pulse data semantic association forward and backward interactive fusion features, and obtaining the recognition result according to the pulse data semantic association forward and backward interactive fusion features.

[0026] In an embodiment of the present application, the step S121, extracting the initial screening abnormal pulse data from the time queue of the pulse data, and dividing the time queue of the pulse data to obtain the time queue of forward pulse data and the time queue of backward pulse data, includes: S1211, preliminarily determining the abnormal pulse data in the time queue of the pulse data to obtain the initial screening abnormal pulse data; S1212, dividing the time queue of the pulse data with the initial screening abnormal pulse data as a node to obtain the time queue of the forward pulse data and the time queue of the backward pulse data.

[0027] Specifically, in step S1211, the abnormal pulse data in the time queue of the pulse data is preliminarily determined to obtain the pre-screening abnormal pulse data. It should be understood that abnormal data often has large volatility and uncertainty. If all the original pulse data are directly analyzed in depth, it may consume a lot of computing resources and time, especially when the amount of data is large. Therefore, in order to remove obviously abnormal data points more quickly, in the technical solution of the present application, the abnormal pulse data in the time queue of the pulse data is preliminarily determined to obtain the pre-screening abnormal pulse data. In this way, the amount of data to be processed later can be reduced, thereby improving the operating efficiency of the entire system. More specifically, when a pulse data time series P={p1, p2, p3,…, pn} is obtained, each data point represents the pulse value measured at different time points. Among them, it is considered that the pulse data will not change drastically in a short time. Therefore, the change rate Δpi=|pi−pi−1| between adjacent data points can be calculated. If the change rate exceeds a predetermined threshold (such as 20 times / minute), the data point is determined to be abnormal. For example, if a pulse value suddenly jumps from 80 to 150 and then returns to 82, it may be due to a measurement error or a sensor problem, and such data points should be marked as abnormal data. Furthermore, in order to reduce misjudgments caused by short-term fluctuations, the sliding window method can be used to smooth the data. A sliding window of size 5 is used to calculate the mean (pi−2, pi−1, pi, pi+1, pi+2) and standard deviation of the data in each window. If the deviation of a data point pi from the window mean exceeds a certain threshold, the point can also be regarded as abnormal data. In this way, obviously abnormal pulse data can be quickly identified and eliminated when the data volume is large, thereby effectively reducing the amount of data that needs to be further processed and improving the computational efficiency of the system.

[0028] Next, in order to more accurately identify and verify the abnormal pulse data points initially screened, in step S1212, the time queue of the pulse data is divided with the initial screening abnormal pulse data as a node to obtain the time queue of the forward pulse data and the time queue of the backward pulse data. In this way, the data change trend before and after the abnormal point is more clearly observed, and the features near the abnormal point are deeply analyzed. In this way, it is determined whether the abnormal point represents a real health problem, or whether it is only caused by measurement error or short-term external interference.

[0029] Figure 4 The present invention is a flowchart of performing semantic association temporal coding on the time queue of the forward pulse data and the time queue of the backward pulse data according to the initial screening abnormal pulse data in the sleep quality monitoring and management method based on the neural network according to the embodiment of the present application to obtain the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features. Figure 4 As shown, in an embodiment of the present application, in the step S122, according to the initial screening abnormal pulse data, the time queue of the forward pulse data and the time queue of the backward pulse data are respectively semantically associated with time series encoding to obtain pulse data semantically associated forward time series features and pulse data semantically associated backward time series features, including: S1221, taking the initial screening abnormal pulse data as a reference semantic, calculating the semantic association between the reference semantic and each forward pulse data in the time queue of the forward pulse data to obtain the time series of the semantic association of the forward pulse data The time queue is formed, and the semantic association between the benchmark semantics and each backward pulse data in the time queue of the backward pulse data is calculated to obtain the time queue of the backward pulse data semantic association; S1222, the time queue of the forward pulse data semantic association and the time queue of the backward pulse data semantic association are respectively encoded with temporal semantic association to obtain a pulse data semantic association forward temporal feature vector as the pulse data semantic association forward temporal feature and a pulse data semantic association backward temporal feature vector as the pulse data semantic association backward temporal feature.

[0030] Specifically, in step S1221, the initial screening abnormal pulse data is used as the benchmark semantics, and the semantic associations between the benchmark semantics and each forward pulse data in the time queue of the forward pulse data are calculated to obtain the time queue of the semantic associations of the forward pulse data, and the semantic associations between the benchmark semantics and each backward pulse data in the time queue of the backward pulse data are calculated to obtain the time queue of the semantic associations of the backward pulse data. It should be understood that, considering that abnormal points are usually accompanied by changes in physiological state, and these changes may have begun to appear before the abnormal points and continue to affect for a period of time afterwards, therefore, in order to capture and mine the correlation between the initial screening abnormal data before and after the initial screening abnormality, so as to help identify which pulse behaviors are closely related in the time series and which pulse behaviors may be isolated or abnormal, in the technical solution of the present application, the initial screening abnormal pulse data is used as the baseline semantics, and the semantic association between the baseline semantics and each forward pulse data in the time queue of the forward pulse data is calculated to obtain the time queue of the semantic association of the forward pulse data, and the semantic association between the baseline semantics and each backward pulse data in the time queue of the backward pulse data is calculated to obtain the time queue of the semantic association of the backward pulse data.

[0031] In particular, in a specific embodiment of the present application, taking the calculation of the semantic association between the initial screening abnormal pulse data as the benchmark semantics and each forward pulse data in the time queue of the forward pulse data to obtain the time queue of the semantic association of the forward pulse data as an example: First, define the window radius The initial value of Usually set to 1 to gradually expand the analysis range. That is, define the window radius To control the dynamic range of the analysis. Then, for each initial window size , calculated in the initial screening of abnormal pulse data The previous time span is Each forward pulse data ,in, The mean value of the cosine similarity with the initial screening abnormal pulse data ; as the semantic relevance of the forward pulse data in the window ,in for and Then gradually increase the window radius Repeat the above steps and record each corresponding semantic association until is the number of forward pulse data in the time queue of the forward pulse data Each calculation obtains a semantic relevance value to obtain the final time queue of the semantic relevance of the forward pulse data. ,In this way, the changes of semantic relevance in different time windows can be evaluated, so as to ,understand the changing trend of pulse data more comprehensively. ,For example, if the semantic relevance gradually decreases, it may indicate the occurrence of ,accelerated heartbeat or other abnormal conditions.

[0032] In an embodiment of the present application, in the step S1222, the time queue of the forward pulse data semantic association and the time queue of the backward pulse data semantic association are respectively subjected to temporal semantic association encoding to obtain a pulse data semantic association forward temporal feature vector as the pulse data semantic association forward temporal feature and a pulse data semantic association backward temporal feature vector as the pulse data semantic association backward temporal feature, including: S1222-1, passing the forward pulse data semantic association time queue through a forward pulse semantic association temporal encoder based on a forward LSTM model to obtain the pulse data semantic association forward temporal feature vector; S1222-2, passing the backward pulse data semantic association time queue through a backward pulse semantic association temporal encoder based on a backward LSTM model to obtain the pulse data semantic association backward temporal feature vector.

[0033] Specifically, in step S1222-1, the time queue of the semantic association of the forward pulse data is passed through a forward pulse semantic association temporal encoder based on a forward LSTM model to obtain the pulse data semantic association forward temporal feature vector. It should be understood that, considering that the semantic association at each time point in the time queue of the semantic association of the forward pulse data reflects the similarity between the pulse data at that moment and the initial screening abnormal pulse data, therefore, in order to be able to further explore and identify whether there are certain specific pulse behavior patterns or features before the initial screening abnormal event occurs, and at the same time capture the trend of pulse data changes over time, in the technical solution of the present application, the time queue of the semantic association of the forward pulse data is passed through a forward pulse semantic association temporal encoder based on a forward LSTM model to utilize the memory unit and gating mechanism inside the forward LSTM model to extract the complex dynamic semantic changes in the forward pulse semantic data, and obtain the pulse data semantic association forward temporal feature vector.

[0034] Specifically, in step S1222-2, the time queue of the semantic association of the backward pulse data is passed through the backward pulse semantic association temporal encoder based on the backward LSTM model to obtain the backward temporal feature vector of the semantic association of the pulse data. It should be understood that in order to further improve the semantic association analysis of pulse data, in addition to encoding the time queue of the semantic association of the forward pulse data, it is also necessary to perform temporal semantic association encoding on the time queue of the semantic association of the backward pulse data. Among them, the time queue of the semantic association of the backward pulse data reflects the semantic similarity between the pulse data and the abnormal data after the initial screening of the abnormal event, so it can help the system capture the potential trend of the pulse data change, especially the pulse behavior pattern after the abnormal event occurs. The backward temporal encoder utilizes the backward LSTM (long short-term memory network) model, which can trace back from the future time point to the current moment, thereby learning the reverse dynamic information in the time series data. The time queue of semantic association of backward pulse data is input into the backward pulse time series encoder based on the backward LSTM model. The structure of the model is similar to the standard LSTM, but it encodes from backward to forward when processing time series. Specifically, the backward LSTM passes the data input of the future moment in reverse through the internal memory unit of the recurrent neural network, and controls the flow and memory of information through gating mechanisms (such as forget gate, input gate and output gate), so that the sequence information of the backward pulse data can be effectively stored and retrieved. Because the backward LSTM can capture the time dependency from the future to the past, it can identify the potential trends and change patterns in the pulse data after the abnormal event occurs.

[0035] Figure 5The present invention is a flowchart of performing dynamic memory mechanism feature interaction on the pulse data semantic association forward time series feature and the pulse data semantic association backward time series feature in the sleep quality monitoring and management method based on neural network according to the embodiment of the present application to obtain the pulse data semantic association forward and backward interactive fusion feature, and obtaining the recognition result according to the pulse data semantic association forward and backward interactive fusion feature. Figure 5 As shown, in an embodiment of the present application, the step S123 performs feature interaction of a dynamic memory mechanism on the pulse data semantically associated forward time series features and the pulse data semantically associated backward time series features to obtain pulse data semantically associated forward and backward interactive fusion features, and obtains a recognition result based on the pulse data semantically associated forward and backward interactive fusion features, including: S1231, performs feature interaction of a dynamic memory mechanism on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector to obtain a pulse data semantically associated forward and backward interactive fusion feature vector as the pulse data semantically associated forward and backward interactive fusion feature; S1232, obtains the recognition result based on the pulse data semantically associated forward and backward interactive fusion feature vector, and the recognition result is used to indicate whether the initial screening abnormal pulse data is determined as abnormal pulse data.

[0036] Specifically, in step S123, the pulse data semantically associated forward time series feature and the pulse data semantically associated backward time series feature are subjected to feature interaction of a dynamic memory mechanism to obtain pulse data semantically associated forward and backward interactive fusion features, and the recognition result is obtained based on the pulse data semantically associated forward and backward interactive fusion features. It should be understood that, considering that the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector respectively express important information about the dynamic semantic changes of pulse data in different time series. Therefore, in order to achieve a deeper level of time series feature fusion, thereby obtaining a richer and more accurate feature representation, so as to more comprehensively understand the relationship between the pulse data at each time point and its previous and next contexts, thereby providing rich data support for the subsequent judgment of abnormal points, in the technical solution of the present application, the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector are subjected to feature interaction of a dynamic memory mechanism to obtain a pulse data semantically associated forward and backward interactive fusion feature vector.

[0037] Figure 6 The present invention is a flowchart of performing a dynamic memory mechanism feature interaction on the pulse data semantic association forward time series feature vector and the pulse data semantic association backward time series feature vector in the sleep quality monitoring and management method based on a neural network according to an embodiment of the present application to obtain a pulse data semantic association forward and backward interactive fusion feature vector. Figure 6As shown, in an embodiment of the present application, the step S1231, performs feature interaction of the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector with a dynamic memory mechanism to obtain a pulse data semantically associated forward and backward interactive fusion feature vector, including: S1231-1, performs collaborative implicit feature extraction on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector to obtain an implicit collaborative coding vector between the pulse data semantically associated forward-backward time series features; S1231-2, performs dynamic attention modulation on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector according to the implicit collaborative coding vector between the pulse data semantically associated forward-backward time series features to obtain the pulse data semantically associated forward and backward interactive fusion feature vector.

[0038] Specifically, in step S1231-1, collaborative implicit feature extraction is performed on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector to obtain an implicit collaborative coding vector between the pulse data semantically associated forward-backward time series features. It should be understood that the collaborative feature extraction network is composed of a point convolution layer and an activation function, which aims to extract high-level abstract features from the original feature vector. Through a carefully designed network architecture, the collaborative feature extraction network can capture the deep-level association between the two feature vectors, the so-called "implicit collaboration". This implicit collaboration is not limited to simple linear combinations, but involves complex nonlinear interaction patterns.

[0039] More specifically, the process of step S1231-1 can be expressed as follows: ;in, and represent the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector, respectively, and They are point-by-point addition, point-by-point subtraction, and point-by-point multiplication. For cascade processing, is the point convolutional encoding operation, is the Leaky ReLU activation function, It is the implicit collaborative encoding vector between the forward-backward time series features of the semantic association of pulse data. Through the convolution operation, the network can identify local association patterns, and the use of activation functions, such as ReLU or Leaky ReLU, ensures that the network can learn nonlinear feature representations, which is crucial for capturing the complex relationships between features.

[0040] In an embodiment of the present application, the step S1231-2, according to the implicit collaborative coding vector between the forward-backward temporal features of the pulse data semantic association, performs dynamic attention modulation on the forward-time feature vector of the pulse data semantic association and the backward-time feature vector of the pulse data semantic association to obtain the forward-backward interactive fusion feature vector of the pulse data semantic association, including: S1231-21, writing the implicit collaborative coding vector between the forward-backward temporal features of the pulse data semantic association into a dynamic memory unit to obtain a forward-backward temporal dynamic key vector of the pulse data semantic association; S1231-22, extracting the forward-backward temporal dynamic key vector of the pulse data semantic association from the dynamic memory unit, and performing dynamic attention modulation on the forward-time feature vector of the pulse data semantic association and the pulse data semantic association; The pulse data semantically associated forward-backward temporal dynamic key vector is feature optimized based on the first converter to obtain an optimized pulse data semantically associated forward-backward temporal feature vector; S1231-23, extracting the pulse data semantically associated forward-backward temporal dynamic key vector from the dynamic memory unit, and performing feature optimization based on the second converter on the pulse data semantically associated backward temporal feature vector and the pulse data semantically associated forward-backward temporal dynamic key vector to obtain an optimized pulse data semantically associated backward temporal feature vector; S1231-24, performing feature linear interaction on the optimized pulse data semantically associated forward temporal feature vector and the optimized pulse data semantically associated backward temporal feature vector to obtain the pulse data semantically associated forward-backward interactive fusion feature vector.

[0041] Specifically, in step S1231-21, the implicit collaborative coding vector between the forward-backward time series features of the pulse data semantic association is written into the dynamic memory unit to obtain the forward-backward time series dynamic key vector of the pulse data semantic association. It should be understood that the dynamic memory unit is a special type of memory structure, and its working principle is similar to the memory unit in the recursive neural network, but with greater flexibility and adaptability. The dynamic memory unit can dynamically adjust its own memory content according to the current input state.

[0042] More specifically, the process of step S1231-21 can be expressed as follows: ;in, is a 1×1 convolution operation, It is the pulse data semantic association forward-backward temporal dynamic key vector. This means that when a new feature vector is input, the dynamic memory unit can adjust the memory according to the newly input data to better adapt to the context requirements of the current scene.

[0043] Specifically, in step S1231-22 and step S1231-23, the pulse data semantically associated forward-backward temporal dynamic key vector is extracted from the dynamic memory unit, and the pulse data semantically associated forward temporal feature vector and the pulse data semantically associated forward-backward temporal dynamic key vector are input into the feature attention modulation module based on the first converter structure to obtain the optimized pulse data semantically associated forward temporal feature vector, and at the same time, the pulse data semantically associated forward-backward temporal dynamic key vector is extracted from the dynamic memory unit, and the pulse data semantically associated backward temporal feature vector and the pulse data semantically associated forward-backward temporal dynamic key vector are input into the feature attention modulation module based on the second converter structure to obtain the optimized pulse data semantically associated backward temporal feature vector. The core of the feature attention modulation module based on the first converter structure and the feature attention modulation module based on the second converter structure is the self-attention mechanism, which allows the model to consider the influence of all other elements when processing each input element, not just adjacent elements. This global perspective helps the model capture long-distance dependencies, which is particularly important for understanding complex data.

[0044] More specifically, the process of the above steps S1231-22 can be expressed by the formula: ;in, is the pulse data semantic association forward query embedding matrix, is the first pulse data semantic association forward bias vector, is matrix multiplication, is the pulse data semantic association forward query vector, is the pulse data semantic association forward value embedding matrix, is the second pulse data semantic association forward bias vector, is the pulse data semantic association forward value vector, yes The transposed vector of yes Length, yes function, It is the semantically associated forward time series feature vector of pulse data after optimization.

[0045] More specifically, the process of the above step S1231-23 can be expressed by the formula: ;in, is the pulse data semantic association backward query embedding matrix, is the first pulse data semantic association backward bias vector, is the pulse data semantic association backward query vector, is the pulse data semantic association backward value embedding matrix, is the semantically associated backward bias vector of the second pulse data, is the pulse data semantic association backward value vector, It is the semantically associated backward time series feature vector of pulse data after optimization.

[0046] In the technical solution of the present application, the feature attention modulation module based on the first converter structure and the feature attention modulation module based on the second converter structure can perform query attention-based feature modulation on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector based on the pulse data semantically associated forward-backward time series dynamic key vector, which is equivalent to using the pulse data semantically associated forward-backward time series dynamic key vector as a common attention space to not only enhance the expression of key information in the feature vector, but also reduce the influence of noise and irrelevant information, while also aligning the optimized pulse data semantically associated forward time series feature vector and the optimized pulse data semantically associated backward time series feature vector in the semantic space, utilizing the subsequent feature interaction and fusion between the two.

[0047] Specifically, in step S1231-24, the optimized pulse data semantic association forward time series feature vector and the optimized pulse data semantic association backward time series feature vector are subjected to feature linear interaction to obtain the pulse data semantic association forward and backward interaction fusion feature vector. It should be understood that the main task of the feature linear interaction network is to further integrate the two optimized feature vectors through linear transformation to generate the final pulse data semantic association forward and backward interaction fusion feature vector.

[0048] More specifically, the process of the above steps S1231-24 can be expressed by the formula: ;in, is the weighting coefficient, It is the pulse data semantic association forward and backward interactive fusion feature vector. In this way, by performing feature linear interaction on the optimized pulse data semantic association forward time series feature vector and the optimized pulse data semantic association backward time series feature vector, the forward and backward information can be combined to fully capture the temporal dynamics and potential laws of the pulse data.

[0049] In an embodiment of the present application, the step S1232, according to the pulse data semantic association forward and backward interactive fusion feature vector, obtains the recognition result, including: passing the pulse data semantic association forward and backward interactive fusion feature vector through a classifier-based pulse abnormal data identifier to obtain the recognition result. That is, the pulse data semantic association forward and backward interactive fusion feature vector obtained by dynamic memory interaction using the pulse data semantic association forward time series feature vector and the pulse data semantic association backward time series feature vector is classified and processed to intelligently determine whether the initial screening abnormal pulse data is determined as abnormal pulse data. In this way, it can automatically adapt to individual physiological differences and environmental changes, reduce the dependence on manual intervention to adjust parameters, and enhance the robustness of the system when encountering unknown situations or new patterns. At the same time, it has a strong dynamic learning ability and can continuously learn new patterns and trends from new data, thereby improving its performance in meeting personalized needs and adapting to dynamic environments.

[0050] In a preferred example, considering that the pulse data semantic association forward time series feature vector and the pulse data semantic association backward time series feature vector respectively represent the short-range-long-range bidirectional time series association characteristics of the forward pulse data semantic association degree and the backward pulse data semantic association degree, when performing attention interaction between features based on the dynamic memory mechanism, the multi-dimensional time series association feature differences caused by the differences in the temporal distribution of data sources in different time domains will cause insufficient correspondence of the dynamic mechanism, resulting in the lack of attention interaction feature instance determination in the pulse data semantic association forward and backward interaction fusion feature vector, thereby affecting the accuracy of the recognition result obtained by the classifier-based pulse abnormality data identifier.

[0051] Based on this, when the pulse data semantic association forward and backward interactive fusion feature vector is passed through the pulse abnormality data identifier based on the classifier, the present application first optimizes the pulse data semantic association forward and backward interactive fusion feature vector, including the steps of: calculating the first pulse data semantic association forward and backward interactive fusion feature vector; and The distance between eigenvalues, e.g. L 2 distance, and take the square root of the distance to obtain the pulse data semantic association forward and backward interactive fusion distance representation matrix, that is: ;in, The first one represents the forward and backward interactive fusion feature vector of the pulse data semantic association eigenvalues, The first one represents the forward and backward interactive fusion feature vector of the pulse data semantic association eigenvalues, represents the forward and backward interactive fusion feature vector of the pulse data semantic association, represents the distance between eigenvalues, The forward and backward interactive fusion distance representation matrix of the pulse data semantic association represents The feature value at each position.

[0052] The pulse data semantic association forward and backward interactive fusion autocorrelation matrix of the pulse data semantic association forward and backward interactive fusion feature vector as a row vector is obtained, that is, ;in, represents the transposed vector of the forward and backward interactive fusion feature vector of the pulse data semantic association, represents matrix multiplication, Represents the forward and backward interactive fusion autocorrelation matrix of the semantic association of pulse data.

[0053] The pulse data semantic association forward and backward interactive fusion feature vector is matrix-multiplied with the pulse data semantic association forward and backward interactive fusion distance representation matrix to obtain the pulse data semantic association forward and backward interactive fusion first-level mapping vector, that is, ;in, represents the pulse data semantic association forward and backward interactive fusion distance representation matrix, Represents the first-level mapping vector of the forward and backward interactive fusion of the semantic association of pulse data.

[0054] The pulse data semantic association forward and backward interactive fusion first-level mapping vector is matrix-multiplied with the pulse data semantic association forward and backward interactive fusion distance representation matrix and the pulse data semantic association forward and backward interactive fusion self-correlation matrix to obtain the pulse data semantic association forward and backward interactive fusion multi-level mapping vector ;in, Represents the pulse data semantic association forward and backward interactive fusion multi-level mapping vector.

[0055] The pulse data semantic association forward and backward interactive fusion multi-level mapping vector is interpolated with the pulse data semantic association forward and backward interactive fusion associated eigenvector composed of the eigenvalues ​​of the pulse data semantic association forward and backward interactive fusion auto-association matrix to obtain an optimized pulse data semantic association forward and backward interactive fusion feature vector, wherein interpolation or zero padding is performed when the eigenvalue is insufficient.

[0056] That is, through the linear target mapping representation of the similarity distance representation matrix of the forward and backward interactive fusion feature vector of the semantic association of the pulse data, the complete similarity instantiation of the self-association of the forward and backward interactive fusion feature vector of the semantic association of the pulse data is instantiated by a secondary target mapping representation based on a multi-level distribution hierarchy, and the negative influence factor of the association mismatch is compensated by the association fusion kernel bias to improve the instance determination degree of the eigenvalue of the forward and backward interactive fusion feature vector of the semantic association of the pulse data under the similarity constraint, that is, the significance degree of the eigenvalue as an instance for classification and regression determination, so as to improve the accuracy of the recognition result obtained by the classifier-based pulse abnormality data identifier of the forward and backward interactive fusion feature vector of the semantic association of the pulse data.

[0057] In summary, step S120 is explained, which extracts and divides the initial screening abnormal pulse data from the time queue of the pulse data, and then performs semantic association temporal coding, so as to intelligently judge whether to determine the initial screening abnormal pulse data as abnormal pulse data according to the dynamic memory interaction representation between the pulse data semantic association forward temporal features and the pulse data semantic association backward temporal features. In this way, it can automatically adapt to individual physiological differences and environmental changes, reduce the dependence on manual intervention to adjust parameters, and enhance the robustness of the system when encountering unknown situations or new patterns. At the same time, it can continuously learn new patterns and trends from new data, and improve performance in meeting personalized needs and adapting to dynamic environments.

[0058] In the above sleep quality monitoring and management method based on neural network, in step S130, the time queue of the pulse data after screening is classified in the intelligent terminal to obtain a classification result, and a mapping relationship is established between the classification result and the sleep stage to obtain a sleep stage mapping result. It should be understood that the pulse data reflects the changes in the physiological state of the human body, and different sleep stages (such as light sleep, deep sleep, rapid eye movement sleep, etc.) are accompanied by different physiological characteristics, including pulse frequency, rhythm and volatility. Therefore, by classifying the pulse data after screening, the pulse characteristics corresponding to different sleep stages can be identified, thereby effectively inferring the sleep state. Among them, the pulse data time queue after screening usually contains cleaner and more effective pulse information, removing obvious outliers or noise. Next, the pulse data are classified using machine learning or deep learning algorithms (such as support vector machines, random forests or neural networks, etc.). The goal of the classification process is to identify the time series pattern of the pulse data, such as the fluctuation frequency and change trend of the pulse, and then classify it into a specific sleep stage. The model uses labeled pulse data and corresponding sleep stage labels during training to learn the implicit relationship between pulse data and sleep stages. Once the pulse data is classified and the classification results are obtained, the next step is to map these classification results with the actual sleep stages. Usually, the division of sleep stages is judged comprehensively based on a variety of physiological indicators (such as electroencephalogram, electrocardiogram, etc.), and pulse data provides auxiliary information. Therefore, by establishing a mapping relationship between the pulse data classification results and the sleep stages, it is possible to associate the pulse signal with a specific sleep state (such as deep sleep, light sleep, wakefulness, etc.). Finally, the sleep stage mapping results output by the smart terminal based on this mapping relationship can provide users with important health information such as their sleep quality and sleep depth.

[0059] In the above-mentioned sleep quality monitoring and management method based on neural network, in step S140, the smart terminal obtains the sleep quality of the monitored object according to the sleep stage mapping result, and provides sleep improvement suggestions for the monitored object according to the sleep quality. It should be understood that the evaluation of sleep quality not only depends on the total sleep time, but also combines multiple factors such as the distribution, depth, and continuity of the sleep stage. Therefore, in the technical solution of the present application, the smart terminal obtains the sleep quality scoring result (for example, good, general, poor, etc.) to provide specific sleep improvement suggestions for the user. For different sleep problems, the terminal can recommend a series of personalized suggestions. For example, if the user has too few deep sleep stages, the system may suggest increasing pre-sleep relaxation activities, avoiding excessive use of electronic devices, or suggesting adjusting the sleep environment (such as temperature, light, noise, etc.). If the user's REM sleep is insufficient, it may be recommended to do appropriate exercise or adjust the evening eating habits. In this way, the smart terminal can not only accurately evaluate the user's sleep quality, but also provide personalized sleep improvement solutions according to actual conditions, help users optimize sleep structure, and improve overall health, so as to achieve better quality of life and health management effects.

[0060] In summary, a sleep quality monitoring and management method based on a neural network according to an embodiment of the present application is explained, which uses a pulse monitoring device to collect pulse data of the monitored object during sleep in real time, forms a time queue and transmits it to the smart terminal. The smart terminal first screens out abnormal data to ensure data accuracy; then, it classifies the screened pulse data, and establishes a mapping relationship between the classification results and the sleep stages to generate a sleep stage mapping result. Finally, the sleep quality of the monitored object is evaluated based on these mapping results, and targeted sleep improvement suggestions are provided. In this way, through precise data processing and analysis, the accuracy of sleep monitoring and the quality of personalized services can be effectively improved.

[0061] Figure 7 FIG. 1 is a system block diagram of a sleep quality monitoring and management system based on a neural network according to an embodiment of the present application. Figure 7As shown, according to the embodiment of the present application, the sleep quality monitoring and management system 100 based on neural network includes: a pulse data time queue extraction module 110, which is used to extract the initial screening abnormal pulse data from the time queue of the pulse data, and divide the time queue of the pulse data to obtain the time queue of forward pulse data and the time queue of backward pulse data; a semantic association time series encoding module 120, which is used to perform semantic association time series encoding on the time queue of the forward pulse data and the time queue of the backward pulse data according to the initial screening abnormal pulse data to obtain the pulse data semantic association forward time series feature and the pulse data semantic association backward time series feature; a dynamic memory mechanism feature interaction module 130, which is used to perform feature interaction of the pulse data semantic association forward time series feature and the pulse data semantic association backward time series feature of the dynamic memory mechanism to obtain the pulse data semantic association forward and backward interaction fusion feature, and obtain the recognition result according to the pulse data semantic association forward and backward interaction fusion feature.

[0062] Here, those skilled in the art can understand that the specific operations of each step in the above-mentioned sleep quality monitoring and management system based on neural network have been referred to above. Figures 1 to 6 The description of the neural network-based sleep quality monitoring and management method has been introduced in detail, and therefore, its repeated description will be omitted.

[0063] As described above, the sleep quality monitoring and management system 100 based on a neural network according to the embodiment of the present application can be implemented in various terminal devices. In one example, the sleep quality monitoring and management system 100 based on a neural network can be integrated into the terminal device as a software module and / or a hardware module. For example, the sleep quality monitoring and management system 100 based on a neural network can be a software module in the operating system of the terminal device, or can be an application developed for the terminal device; of course, the sleep quality monitoring and management system 100 based on a neural network can also be one of the many hardware modules of the terminal device.

[0064] Alternatively, in another example, the neural network-based sleep quality monitoring and management system 100 and the terminal device may also be separate devices, and the neural network-based sleep quality monitoring and management system 100 may be connected to the terminal device via a wired and / or wireless network, and transmit interactive information in accordance with an agreed data format.

[0065] In the embodiments provided by the present invention, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the modules is only a logical function division, and there may be other division methods in actual implementation.

[0066] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0067] In addition, each functional module in each embodiment of the present invention may be integrated into one processing unit, each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of hardware plus software functional modules.

[0068] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0069] Therefore, no matter from which point of view, the embodiments should be regarded as illustrative and non-restrictive, and the scope of the present invention is limited by the appended claims rather than the above description, so it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any attached figure mark in the claims should not be regarded as limiting the claims involved.

[0070] In addition, it is clear that the word "comprising" does not exclude other units or steps, and the singular does not exclude the plural. Multiple units or devices stated in the system claim can also be implemented by one unit or device through software or hardware. The second and other words are used to indicate names, but not to indicate any particular order.

[0071] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. 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 solution of the present invention can be modified or replaced by equivalents without departing from the spirit of the technical solution of the present invention.

Claims

1. A sleep quality monitoring and management method based on a neural network, comprising: Using a pulse monitoring device to collect pulse data of a monitored subject in real time during sleep to obtain a time queue of the pulse data, and transmitting the time queue of the pulse data to an intelligent terminal; at the intelligent terminal, filtering abnormal data from the time queue of the pulse data to obtain a time queue of filtered pulse data; at the intelligent terminal, classifying the time queue of filtered pulse data to obtain a classification result, and establishing a mapping relationship between the classification result and the sleep stage to obtain a sleep stage mapping result; In the intelligent terminal, the sleep quality of the monitored object is obtained according to the sleep stage mapping result, and sleep improvement suggestions are provided to the monitored object according to the sleep quality; it is characterized in that the time queue of the pulse data is screened for abnormal data to obtain the screened pulse data, including: extracting the initial screening abnormal pulse data from the time queue of the pulse data, and dividing the time queue of the pulse data to obtain the time queue of forward pulse data and the time queue of backward pulse data; according to the initial screening abnormal pulse data, semantically associated temporal encoding is performed on the time queue of the forward pulse data and the time queue of the backward pulse data respectively to obtain the pulse data semantically associated forward temporal features and the pulse data semantically associated backward temporal features; the pulse data semantically associated forward temporal features and the pulse data semantically associated backward temporal features are subjected to feature interaction of a dynamic memory mechanism to obtain the pulse data semantically associated forward and backward interactive fusion features, and the recognition result is obtained according to the pulse data semantically associated forward and backward interactive fusion features.

2. The sleep quality monitoring and management method based on neural network according to claim 1, characterized in that: Extracting preliminary screening abnormal pulse data from the time queue of the pulse data, and dividing the time queue of the pulse data to obtain a time queue of forward pulse data and a time queue of backward pulse data, including: preliminarily determining the abnormal pulse data in the time queue of the pulse data to obtain the preliminary screening abnormal pulse data; dividing the time queue of the pulse data with the preliminary screening abnormal pulse data as a node to obtain the time queue of the forward pulse data and the time queue of the backward pulse data.

3. The sleep quality monitoring and management method based on neural network according to claim 2 is characterized in that: According to the initial screening abnormal pulse data, the time queue of the forward pulse data and the time queue of the backward pulse data are respectively semantically associated with time series coding to obtain pulse data semantically associated forward time series features and pulse data semantically associated backward time series features, including: taking the initial screening abnormal pulse data as the reference semantics, calculating the semantic association between the reference semantics and each forward pulse data in the time queue of the forward pulse data to obtain the time queue of the forward pulse data semantic association, and calculating the semantic association between the reference semantics and each backward pulse data in the time queue of the backward pulse data to obtain the time queue of the backward pulse data semantic association; performing time series semantic association coding on the time queue of the forward pulse data semantic association and the time queue of the backward pulse data semantic association to obtain pulse data semantically associated forward time series feature vector as the pulse data semantically associated forward time series feature and pulse data semantically associated backward time series feature vector as the pulse data semantically associated backward time series feature.

4. The sleep quality monitoring and management method based on neural network according to claim 3 is characterized in that: The time queue of the forward pulse data semantic association and the time queue of the backward pulse data semantic association are respectively subjected to temporal semantic association encoding to obtain a pulse data semantic association forward temporal feature vector as the pulse data semantic association forward temporal feature and a pulse data semantic association backward temporal feature vector as the pulse data semantic association backward temporal feature, including: passing the forward pulse data semantic association time queue through a forward pulse semantic association temporal encoder based on a forward LSTM model to obtain the pulse data semantic association forward temporal feature vector; passing the backward pulse data semantic association time queue through a backward pulse semantic association temporal encoder based on a backward LSTM model to obtain the pulse data semantic association backward temporal feature vector.

5. The sleep quality monitoring and management method based on neural network according to claim 4 is characterized in that: The pulse data semantically associated forward time series features and the pulse data semantically associated backward time series features are subjected to feature interaction of a dynamic memory mechanism to obtain pulse data semantically associated forward and backward interactive fusion features, and a recognition result is obtained based on the pulse data semantically associated forward and backward interactive fusion features, including: the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector are subjected to feature interaction of a dynamic memory mechanism to obtain pulse data semantically associated forward and backward interactive fusion feature vector as the pulse data semantically associated forward and backward interactive fusion features; the recognition result is obtained based on the pulse data semantically associated forward and backward interactive fusion feature vector, and the recognition result is used to indicate whether the initially screened abnormal pulse data is determined as abnormal pulse data.

6. The sleep quality monitoring and management method based on neural network according to claim 5 is characterized in that: The pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector are subjected to feature interaction of a dynamic memory mechanism to obtain a pulse data semantically associated forward and backward interactive fusion feature vector, including: performing collaborative implicit feature extraction on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector to obtain an implicit collaborative coding vector between the pulse data semantically associated forward-backward time series features; and performing dynamic attention modulation on the pulse data semantically associated forward time series feature vector and the pulse data semantically associated backward time series feature vector according to the implicit collaborative coding vector between the pulse data semantically associated forward-backward time series features to obtain the pulse data semantically associated forward and backward interactive fusion feature vector.

7. The sleep quality monitoring and management method based on neural network according to claim 6 is characterized in that: According to the implicit collaborative coding vector between the forward-backward temporal features of the pulse data semantic association, the forward-time feature vector of the pulse data semantic association and the backward-time feature vector of the pulse data semantic association are dynamically attention modulated to obtain the forward-backward interactive fusion feature vector of the pulse data semantic association, including: writing the implicit collaborative coding vector between the forward-backward temporal features of the pulse data semantic association into a dynamic memory unit to obtain a forward-backward temporal dynamic key vector of the pulse data semantic association; extracting the forward-backward temporal dynamic key vector of the pulse data semantic association from the dynamic memory unit, and modulating the forward-time feature vector of the pulse data semantic association and the forward-backward temporal dynamic key vector of the pulse data semantic association. The pulse data semantic association forward-backward time series dynamic key vector is feature optimized based on the first converter to obtain an optimized pulse data semantic association forward time series feature vector; the pulse data semantic association forward-backward time series dynamic key vector is extracted from the dynamic memory unit, and the pulse data semantic association backward time series feature vector and the pulse data semantic association forward-backward time series dynamic key vector are feature optimized based on the second converter to obtain an optimized pulse data semantic association backward time series feature vector; the optimized pulse data semantic association forward time series feature vector and the optimized pulse data semantic association backward time series feature vector are feature linearly interacted to obtain the pulse data semantic association forward-backward interactive fusion feature vector.

8. The sleep quality monitoring and management method based on neural network according to claim 7 is characterized in that: The recognition result is obtained according to the pulse data semantically associated forward and backward interactive fusion feature vector, including: passing the pulse data semantically associated forward and backward interactive fusion feature vector through a classifier-based pulse abnormality data identifier to obtain the recognition result.

9. A sleep quality monitoring and management system based on neural network, characterized in that: include: A pulse data time queue extraction module, used to extract primary screening abnormal pulse data from the pulse data time queue, and divide the pulse data time queue to obtain a forward pulse data time queue and a backward pulse data time queue; A semantic association temporal coding module, for performing semantic association temporal coding on the time queue of the forward pulse data and the time queue of the backward pulse data respectively according to the primary screening abnormal pulse data to obtain a pulse data semantic association forward temporal feature and a pulse data semantic association backward temporal feature; The dynamic memory mechanism feature interaction module is used to perform feature interaction of the pulse data semantic association forward time series feature and the pulse data semantic association backward time series feature to obtain the pulse data semantic association forward and backward interaction fusion feature, and obtain the recognition result based on the pulse data semantic association forward and backward interaction fusion feature.

10. The sleep quality monitoring and management system based on neural network according to claim 9, characterized in that: The pulse data time queue extraction module includes: preliminarily determining the abnormal pulse data in the time queue of the pulse data to obtain the initial screening abnormal pulse data; dividing the time queue of the pulse data with the initial screening abnormal pulse data as a node to obtain the time queue of the forward pulse data and the time queue of the backward pulse data.

Images