A method and system for monitoring source patient signal data analysis

Abstract

The application relates to the field of data processing, and particularly provides a monitoring source patient signal data analysis method and system. The method comprises the following steps: obtaining target patient behavior signals matched with to-be-detected behavior signals, wherein the target patient behavior signals cover behavior signals of different categories; obtaining matching degrees between the behavior signals of different categories covered by the target patient behavior signals and the to-be-detected behavior signals based on influence factors of the behavior signals of different categories; screening selected category behavior signals from the behavior signals of different categories based on the matching degrees between the behavior signals of different categories and the to-be-detected behavior signals; and performing highlight display on signal segments in the selected category behavior signals matched with the to-be-detected behavior signals. The application can simplify the highlight display signals, highlight the signals with more analysis significance, and help more accurate recognition of signal behavior patterns.

Description

Technical Field

[0001] This application relates to the field of data processing, and more specifically, to a method and system for analyzing patient signal data from a monitoring source. Background Technology

[0002] In the field of medical data analytics, with the widespread adoption of wearable devices and telemedicine services, collecting and analyzing patient behavior data has become a crucial means of improving the quality of healthcare services and personalizing treatment. Patient behavior data typically includes various categories of signals, such as heart rate, blood pressure, step count, and sleep quality. This data is of great significance for understanding a patient's health status, predicting disease risk, and developing personalized treatment plans.

[0003] In practical applications, effectively extracting key information related to specific health events or symptoms from massive and complex patient behavioral data remains a significant challenge. Traditional data analysis methods often treat all data in a one-size-fits-all manner, ignoring the differences between different categories of behavioral signals and their varying contributions to health events. This approach may not only lead to the omission of crucial information but also obscure potential patterns and trends within the data due to homogenization. Furthermore, existing data analysis tools often lack the highlighting of key signals when presenting analysis results, making it difficult for healthcare professionals to quickly identify data points highly correlated with the behavioral signals being analyzed. This not only increases the complexity and time cost of data analysis but may also affect the accuracy and timeliness of medical decisions. Summary of the Invention

[0004] The purpose of this application is to provide a method and system for analyzing patient signal data from a monitoring source.

[0005] This application is implemented as follows:

[0006] In a first aspect, this application provides a method for analyzing patient signal data from a monitoring source, the method comprising:

[0007] Acquire target patient behavior signals that are compatible with the behavior signals to be detected, wherein the target patient behavior signals cover different categories of behavior signals;

[0008] Based on the influence factors of the different categories of behavioral signals, the matching degree between the different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected is obtained;

[0009] Based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected, a selected category of behavioral signals is obtained by filtering from the different categories of behavioral signals;

[0010] The signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are highlighted and displayed.

[0011] The step of highlighting and displaying the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected includes:

[0012] Multiple signal segments to be emphasized are retrieved from the selected category of behavioral signals and are adapted to the behavioral signal to be detected;

[0013] The continuity evaluation parameters among the multiple signal segments to be emphasized are determined by the signal coordinates of the multiple signal segments to be emphasized in the behavioral signal of the selected category;

[0014] When the continuity evaluation parameters among the multiple signal segments to be emphasized reach the preset standard, the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are emphasized and displayed.

[0015] Optionally, obtaining the matching degree between the different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected based on the influence factors of the different categories of behavioral signals includes:

[0016] The signal to be detected is segmented into signal segments to obtain the signal segments to be detected;

[0017] Based on the signal segment to be detected, the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected is obtained;

[0018] Based on the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected, and the influence factor of the different categories of behavioral signals, the matching degree between the different categories of behavioral signals and the behavioral signal to be detected is obtained.

[0019] Optionally, obtaining the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected based on the signal segment to be detected includes:

[0020] Target signal segments that match the signal segment to be detected are retrieved from the different categories of behavioral signals;

[0021] Based on the priority of the signal segment to be detected and the priority of the target signal segment that matches the signal segment to be detected among the different categories of behavioral signals, the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected is determined.

[0022] Optionally, the method may include multiple target patient behavioral signals, and may further include:

[0023] Based on the adaptation mechanism between the behavioral signal to be detected and the behavioral signal of each target patient, the influence factor of each target patient's behavioral signal is determined;

[0024] Based on the influence factor of each target patient behavior signal and the matching degree between the different categories of behavior signals covered by each target patient behavior signal and the behavior signal to be detected, the matching degree between each target patient behavior signal and the behavior signal to be detected is obtained;

[0025] Based on the matching degree between each target patient behavior signal and the behavior signal to be detected, multiple target patient behavior signals are listed in sequence.

[0026] The acquisition of the target patient behavioral signal that matches the behavioral signal to be detected includes:

[0027] The signal to be detected is segmented into signal segments to obtain a set of signal segments to be detected;

[0028] The signal segments in the set of signal segments to be detected are divided into clusters to obtain multiple signal segment clusters;

[0029] Target patient behavior signals that are adapted to the multiple signal segment clusters are acquired respectively, with each signal segment cluster corresponding to one adaptation mechanism.

[0030] Optionally, the step of segmenting the signal to be detected behavior signal to obtain a set of signal segments to be detected includes:

[0031] The behavior patterns of the detected behavior signals are classified to obtain behavior pattern labels for the detected behavior signals;

[0032] The signal to be detected is segmented, and the set of signal segments to be detected is generated based on the behavior pattern label and the signal segments obtained from the signal segmentation.

[0033] Optionally, the step of filtering out a selected category of behavioral signals from the different categories of behavioral signals based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected includes:

[0034] The target category of behavioral signals is obtained by filtering from the different categories of behavioral signals, as well as behavioral signals of unselected categories other than the target category;

[0035] Based on the matching degree between the different categories of behavioral signals and the behavioral signals to be detected, behavioral signals of undetermined categories are filtered from the behavioral signals of unselected categories.

[0036] The behavioral signals of the undetermined category and the behavioral signals of the target category are used as the behavioral signals of the selected category;

[0037] The method of highlighting and displaying signal segments in the selected category of behavioral signals that match the behavioral signal to be detected also includes:

[0038] The target category signal segments are obtained by filtering from the set of signal segments to be detected, and the set of signal segments to be detected includes signal segments obtained by segmenting the signal segments of the behavior signal to be detected.

[0039] The signal segments in the selected category of behavioral signals that match the signal segments of the target category are highlighted and displayed.

[0040] Optionally, the step of filtering behavioral signals of unselected categories from the behavioral signals based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected includes:

[0041] Based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected, the original undetermined category of behavioral signals with a matching degree higher than the matching degree threshold is selected from the behavioral signals of unselected categories.

[0042] The behavioral signals of the original undetermined categories corresponding to the top K matching degrees in descending order are taken as the behavioral signals of the undetermined categories, where K≥1.

[0043] Optionally, the step of highlighting and displaying the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected further includes:

[0044] Identify unselected category signal segments in the set of signal segments to be detected, other than those of the target category.

[0045] A candidate signal segment set is obtained by retrieving signal segments that match the unselected category signal segments from the selected category of behavioral signals.

[0046] G signal segments are selected from the candidate signal segment set and highlighted for display, where G≥1.

[0047] Optionally, retrieving multiple signal segments to be emphasized from the selected category of behavioral signals that are adapted to the behavioral signal to be detected includes:

[0048] In the detected behavior signal, obtain the first signal segment and the second signal segment whose signal coordinates are adjacent;

[0049] A third signal segment that matches the first signal segment and a fourth signal segment that matches the second signal segment are retrieved from the selected category of behavioral signals.

[0050] The step of determining the continuity evaluation parameters among the plurality of signal segments to be emphasized by using the signal coordinates of the plurality of signal segments to be emphasized in the behavioral signal of the selected category includes:

[0051] Based on the signal coordinates of the third signal segment and the fourth signal segment in the selected category of behavioral signals, a continuity evaluation parameter between the third signal segment and the fourth signal segment is determined.

[0052] When the continuity evaluation parameters among the plurality of signal segments to be emphasized reach a preset standard, the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are emphasized and displayed, including:

[0053] When the continuity evaluation parameter between the third signal segment and the fourth signal segment reaches the preset standard, a signal segment that matches the behavior signal to be detected is selected from the selected category of behavior signals, and the selected signal segment is highlighted and displayed.

[0054] The step of determining the continuity evaluation parameters between the third signal segment and the fourth signal segment based on the signal coordinates of the third signal segment and the fourth signal segment in the selected category of behavioral signals includes:

[0055] Based on the signal coordinates of the third signal segment and the fourth signal segment in the selected category of behavioral signals, the signal spacing between the third signal segment and the fourth signal segment is determined.

[0056] The continuity evaluation parameters between the third signal segment and the fourth signal segment are determined based on the signal coordinate reference of the third signal segment and the fourth signal segment, as well as the spacing between the third signal segment and the fourth signal segment.

[0057] In a second aspect, this application provides a computer system comprising: one or more processors; a memory; and one or more computer programs; wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more computer programs are executed by the processors, they implement the method described above.

[0058] The beneficial effects of this application are as follows: This application determines the matching degree between different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected by combining the influencing factors of different types of behavioral signals. The obtained matching degree can characterize the contribution of different categories of behavioral signals to the behavioral signals to be detected, thus improving the accuracy of the obtained matching degree. By using the matching degree between different categories of behavioral signals and the behavioral signals to be detected, selected categories of behavioral signals are screened from the different categories of behavioral signals covered by the target patient's behavioral signals. The signal segments in the selected categories of behavioral signals that match the behavioral signals to be detected are highlighted and displayed. This can screen out behavioral signals that match the behavioral signals to be detected more closely from the target patient's behavioral signals, improve the correlation between the highlighted signals and the behavioral signals to be detected, and prevent the homogeneous highlighting and display of different categories of behavioral signals in the target patient's behavioral signals, which would result in the core signal intervals not being highlighted and displayed. In other words, this application can streamline the highlighted signals and highlight the signals that are more meaningful for analysis, helping to more accurately identify signal behavior patterns. Attached Figure Description

[0059] Figure 1 This is a flowchart of a method for analyzing patient signal data from a monitoring source, provided in an embodiment of this application.

[0060] Figure 2 This is a schematic diagram of the composition of a computer system provided in an embodiment of this application. Detailed Implementation

[0061] The embodiments of this application are described below with reference to the accompanying drawings. The terminology used in the implementation section of this application is for explaining specific embodiments only and is not intended to limit the scope of this application.

[0062] In this embodiment, the execution entity of the monitoring source patient signal data analysis method is a computer system, including but not limited to servers, personal computers, laptops, tablets, and smartphones. The computer system can run independently to implement this application, or it can be connected to a network and implement this application through interactive operation with other computer systems on the network. The network where the computer system is located includes, but is not limited to, the Internet, wide area networks, metropolitan area networks, local area networks, and VPN networks.

[0063] like Figure 1 As shown, the method includes:

[0064] Step S110: Obtain the target patient behavior signal that matches the behavior signal to be detected, which includes different categories of behavior signals.

[0065] In step S110, the computer system acquires the target patient's behavioral signal that matches the behavioral signal to be detected. First, the computer system clarifies the specific content of the behavioral signal to be detected. The behavioral signal to be detected may be one or more specific physiological indicators, behavioral activities, or medical events, such as a patient's heart rate changes, daily activity levels, medication use, etc. Taking heart rate changes as an example, suppose we want to analyze the reason for a patient's abnormally elevated heart rate over a specific time period; then the heart rate data is the behavioral signal to be detected.

[0066] Next, the computer system searches its database for target patient behavioral signals that are relevant to the behavioral signal being detected. This data typically originates from patient health monitoring devices, electronic medical record systems, or other healthcare information systems. The target patient behavioral signals not only include information directly related to the behavioral signal being detected (such as the patient's electrocardiogram record) but may also encompass other factors that may influence the behavioral signal, such as the patient's age, gender, weight, medical history, and lifestyle habits. This information is organized into different categories of behavioral signals to facilitate subsequent analysis and processing.

[0067] To illustrate more specifically, the hypothetical behavioral signal to be detected is a sudden increase in the patient's heart rate at a certain time. The computer system searches its database for the patient's electrocardiogram (ECG) records as directly relevant target patient behavioral signals. Simultaneously, it also retrieves indirectly relevant behavioral signals such as the patient's age (physiological factors that may affect heart rate changes), gender (physiological differences that may affect heart rate changes), weight (which may affect cardiac workload and heart rate), medical history (such as heart disease, hypertension, and other diseases that may affect heart rate), and lifestyle habits (such as whether the patient exercises regularly, smokes, drinks alcohol, and other daily behaviors that may affect heart rate).

[0068] In acquiring target patient behavioral signals, computer systems can utilize machine learning models to optimize the search and filtering process. For example, it can train a classification model that automatically identifies and retrieves the most relevant target patient behavioral signal categories based on the input behavioral signal to be detected (such as heart rate changes). This model can be trained using historical data, where the input features include the type and value of the behavioral signal to be detected, and the output is a probability distribution representing the degree of association between each behavioral signal category and the behavioral signal to be detected.

[0069] When training a classification model, computer systems can use supervised learning algorithms such as Support Vector Machines (SVM), Random Forests, or Gradient Boosting Machines. These algorithms learn a decision boundary or function from labeled training data to distinguish the degree of association between different categories of behavioral signals and the behavioral signal to be detected. During training, the computer system continuously adjusts the model parameters to minimize prediction error and improve classification accuracy. After training, the model is applied to a real-world data retrieval task. For a new behavioral signal to be detected, the computer system inputs it into the model, and the model outputs a probability distribution indicating the degree of association between each behavioral signal category and the behavioral signal to be detected. Based on this probability distribution, the computer system can select the several behavioral signal categories with the highest degree of association as the target patient behavioral signals for subsequent analysis.

[0070] In addition to classification models, computer systems can utilize clustering algorithms to further organize target patient behavioral signals. Clustering algorithms group similar behavioral signals together, making it easier for analysts to identify patterns and trends in the data. For example, the K-means clustering algorithm can divide behavioral signals into K clusters based on their numerical characteristics, with each cluster containing numerically similar behavioral signals. In this way, the computer system can organize the target patient behavioral signals into different clusters, each cluster representing a specific behavioral pattern or patient group.

[0071] Step S120: Based on the influencing factors of different categories of behavioral signals, obtain the matching degree between the different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected.

[0072] In step S120, the computer system calculates the matching degree between different categories of behavioral signals in the target patient's behavioral signals and the behavioral signals to be detected. This process is based on the influence factors (i.e., weights) of different categories of behavioral signals, aiming to assess the contribution of each behavioral signal to the behavioral signals to be detected, thereby selecting the most relevant behavioral signals for analysis. First, the computer system determines the influence factors of different categories of behavioral signals. These influence factors, also known as weights, reflect the importance of each behavioral signal in a specific analytical scenario. The determination of influence factors can be based on various factors, including but not limited to statistical analysis of historical data, expert opinions, and prediction results from machine learning models.

[0073] Taking heart rate and blood pressure as examples, suppose we are analyzing a patient's abnormal heart rate. In this case, both heart rate and blood pressure signals are important behavioral signal categories. However, their influence on abnormal heart rate phenomena may differ. Statistical analysis of historical data reveals a higher correlation between heart rate signals and abnormal heart rate phenomena, thus allowing for a higher weighting. Conversely, while blood pressure signals are also related to heart health, they may not be a primary factor in the current analytical context, thus warranting a lower weighting.

[0074] Besides statistical analysis, influencing factors can also be determined using machine learning models. For example, a regression model can be trained that uses each behavioral signal as input features and the behavioral signal to be detected as the target variable, optimizing the model parameters by minimizing the prediction error. During training, the model automatically learns the degree of influence of each input feature (i.e., behavioral signal) on the target variable (i.e., the behavioral signal to be detected) and represents these influences as weights. After training, this model can be used to calculate the matching degree between each behavioral signal in a new dataset and the behavioral signal to be detected.

[0075] Once the influencing factors of each behavioral signal are determined, the computer system can proceed to the matching degree calculation stage. This process typically involves segmenting the target patient's behavioral signal and calculating the matching degree between each segment and the behavioral signal to be detected. The matching degree can be calculated using various methods, such as cosine similarity, Euclidean distance, and Pearson correlation coefficient. Each of these methods has its advantages and disadvantages, and the choice of method depends on the specific application scenario and data characteristics.

[0076] Taking cosine similarity as an example, suppose we have two vectors A and B, representing the target behavior signal and a certain behavior signal segment, respectively. The formula for calculating cosine similarity is:

[0077] ;

[0078] in, This represents the dot product of vectors A and B. Let A and B represent the magnitudes of vectors A and B, respectively. The cosine similarity value ranges from -1 to 1, with values ​​closer to 1 indicating greater similarity between the two vectors.

[0079] In practical applications, the target behavior signal and the behavior signal segment can be represented as feature vectors, where each feature corresponds to a value of a behavior signal. Then, the cosine similarity between them is calculated using the formula above, serving as a measure of the matching degree. It is important to note that since an influence factor (weight) has been determined for each behavior signal category, the corresponding feature values ​​need to be weighted when calculating the dot product.

[0080] Specifically, assume there are three categories of behavioral signals: heart rate, blood pressure, and activity level, with influence factors of 0.6, 0.3, and 0.1, respectively. Also, assume the behavioral signal to be detected and the behavioral signal segment can be represented by the following feature vectors: .

[0081] When calculating cosine similarity, the feature values ​​of the behavioral signal segments need to be weighted first: Then, cosine similarity is calculated using the weighted behavioral signal segment and the behavioral signal to be detected:

[0082] ;

[0083] By calculating the cosine similarity between each behavior signal segment and the behavior signal to be detected, a matching degree matrix can be obtained, where each element represents the degree of matching between the corresponding behavior signal segment and the behavior signal to be detected. Finally, the computer system can use this matching degree matrix to select the most relevant behavior signal segments for subsequent analysis.

[0084] Step S130: Based on the matching degree between different categories of behavioral signals and the behavioral signals to be detected, select behavioral signals of the selected category from the different categories of behavioral signals.

[0085] In step S130, the computer system selects the behavior signal category that is most relevant to the behavior signal to be detected and has the highest matching degree from among many behavior signal categories, based on the matching degree between the different categories of behavior signals and the behavior signal to be detected calculated in step S120, as the selected category for subsequent analysis.

[0086] First, the computer system sorts the matching degrees calculated in step S120. Assume a matching degree matrix is ​​already obtained, where each row represents a behavioral signal category, each column represents a behavioral signal segment to be detected, and the elements in the matrix are the matching degrees between the corresponding behavioral signal category and the behavioral signal segment to be detected. The computer system can sort this matrix by row (or column), placing the behavioral signal category with the highest matching degree first. Taking heart rate, blood pressure, and activity level as examples, assume that the matching degrees between these three behavioral signal categories and a certain behavioral signal segment to be detected (such as abnormal heart rate over a certain period of time) have already been calculated.

[0087] Next, the computer system sets a filtering threshold or selects a certain number of behavioral signal categories as selected categories based on the sorted matching degree matrix. The filtering threshold can be set according to the actual analysis needs and data characteristics. For example, it can select all behavioral signal categories with a matching degree higher than a certain specific value, or select the top N behavioral signal categories with the highest matching degree (N is a preset integer).

[0088] Continuing with the above example, assuming a filtering threshold of 0.7, only behavioral signal categories with a match rate higher than 0.7 will be selected as the chosen category. In this case, only the heart rate signal meets the condition, therefore the heart rate signal will be selected as the chosen category for subsequent analysis.

[0089] Furthermore, if the top N behavioral signal categories with the highest matching degree are selected as the chosen categories, the computer system will select categories sequentially from highest to lowest matching degree until the number of selected behavioral signal categories reaches N. For example, if the top two behavioral signal categories with the highest matching degree are selected, then heart rate and blood pressure signals will be selected as the chosen categories.

[0090] When setting screening thresholds or selecting the number of behavioral signal categories, computer systems can consider the impact of factors such as data sparsity and noise interference on the matching degree. If there are a large number of missing or outlier values ​​in the data, these values ​​can reduce the accuracy of the matching degree. To address this, computer systems can employ data preprocessing techniques (such as interpolation, smoothing, and denoising) to improve the integrity and accuracy of the data, thereby ensuring the reliability of the matching degree calculation.

[0091] Furthermore, in certain complex scenarios, computer systems can utilize machine learning models to assist in selecting behavioral signals of chosen categories. For example, a classification model can be trained to predict which behavioral signal categories are most relevant to the behavioral signal to be detected. This model can be trained based on historical data, where input features include the matching degree between each behavioral signal category and the behavioral signal to be detected, statistical characteristics of the behavioral signal categories, etc., and the output is a binary label (indicating whether the behavioral signal category is selected as the chosen category). After the model is trained, the computer system can input new data into the model for prediction and select behavioral signals of chosen categories based on the prediction results.

[0092] Step S140: Highlight and display the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected.

[0093] In step S140, the computer system highlights and displays the signal segments from the selected category of behavioral signals that match the behavioral signal to be detected. This step aims to help users (such as healthcare professionals) quickly identify the most relevant and valuable behavioral signal segments for the target analysis, thereby obtaining key information more efficiently and making accurate judgments.

[0094] First, the computer system determines which signal segments are compatible with the behavioral signal to be detected. This typically involves further subdividing the selected category of behavioral signals into smaller segments and evaluating the correlation between each segment and the behavioral signal to be detected. This correlation can be measured using various metrics, such as temporal synchronicity, numerical similarity, and trend consistency. For example, in heart rate monitoring, if the heart rate variability pattern of a signal segment matches the behavioral signal to be detected (such as a patient suddenly experiencing palpitations), then that signal segment can be considered compatible.

[0095] In determining the appropriate signal segment, computer systems can utilize advanced signal processing techniques, such as filtering, smoothing, and feature extraction, to improve signal quality and reduce noise interference. These techniques help to more accurately identify key features and trends in the signal.

[0096] Once the appropriate signal segments are determined, the computer system considers how to emphasize them. The purpose of emphasis is to make these signal segments stand out more in the user interface so that users can immediately recognize their importance. There are many ways to achieve emphasis, including but not limited to changing visual attributes such as the color, thickness, and flashing frequency of the signal segments, or using sensory feedback such as sound and vibration to attract the user's attention.

[0097] In practice, computer systems can set different emphasis levels based on the importance of signal segments. For example, key signal segments that are highly correlated with the detected behavioral signal can be highlighted with bright colors (such as red) and thicker lines; while signal segments with lower correlation but still some reference value can be highlighted with a softer emphasis.

[0098] Furthermore, to enhance the presentation, the computer system can combine the adapted signal segment with its contextual information. Contextual information can include the signal segment's timestamp, acquisition location, patient status, and other relevant data. By presenting this information graphically or textually on the user interface, users can quickly understand the specific content of the signal segment and better comprehend its background and conditions.

[0099] In some cases, computer systems can also leverage machine learning models to assist in the emphasis presentation process. For example, a classification model can be trained to predict which signal segments are most likely associated with the signal of the behavior to be detected, and the emphasis presentation strategy can be dynamically adjusted based on the prediction results. Such models can be built on historical data and expert knowledge, and their predictive accuracy can be continuously improved by updating data and algorithms.

[0100] In terms of the specific implementation of the display, computer systems can employ various technical means. For example, interactive chart libraries (such as D3.js, ECharts, etc.) can be used to create dynamic and interactive signal display interfaces, allowing users to view the details of signal segments through zooming, panning, and other operations. Simultaneously, natural language processing techniques can be combined to generate text descriptions and explanations of signal segments, helping users better understand the meaning and importance represented by the signal segments.

[0101] Finally, it's important to emphasize that presentation isn't just about aesthetics or attracting user attention; more importantly, it's about ensuring users can accurately and quickly obtain key information and make informed decisions. Therefore, when designing presentation schemes, computer systems must fully consider user needs and habits to ensure that the presentation is both intuitive and easy to understand, as well as efficient and practical.

[0102] In summary, by comprehensively utilizing technologies such as signal processing, visualization, and machine learning, step S140 enables the computer system to provide users with a more intuitive, accurate, and efficient way of displaying information, helping users to better understand and cope with complex medical data analysis tasks.

[0103] In one implementation, step S140 emphasizes and displays the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected, including:

[0104] Step S141: Retrieve multiple signal segments to be emphasized from the selected category of behavioral signals that are compatible with the behavioral signal to be detected;

[0105] Step S142: Determine the continuity evaluation parameters between multiple signal segments to be emphasized by using the signal coordinates of multiple signal segments to be emphasized in the behavioral signals of the selected category;

[0106] Step S143: When the continuity evaluation parameters between multiple signal segments to be emphasized reach the preset standard, the signal segments in the selected category of behavioral signals that match the behavioral signals to be detected are emphasized and displayed.

[0107] In step S141, the computer system identifies signal segments from the selected category of behavioral signals that highly match the behavioral signal to be detected. Fit is typically assessed based on the similarity between the signal segment and the behavioral signal to be detected in terms of value, trend, and pattern. To achieve this, the computer system can employ various methods, including but not limited to signal matching algorithms and pattern recognition techniques. For example, suppose the electrocardiogram data of a heart patient is being analyzed, where the selected category of behavioral signal is the heart rate signal. The behavioral signal to be detected is the heart rate change during a heart attack in this patient. The computer system retrieves signal segments from the heart rate signal that are similar in pattern to the heart rate change during this heart attack as signal segments to be emphasized.

[0108] First, the computer system divides the heart rate signal into multiple shorter signal segments, each containing continuous heart rate data points. For each segment, the system extracts a series of features, such as mean, standard deviation, kurtosis, and skewness, which summarize the segment's main statistical characteristics. Then, the system calculates the similarity between each segment and the target behavioral signal. This can be achieved by calculating the distance (e.g., Euclidean distance) or correlation (e.g., Pearson correlation coefficient) between the segment's features and the target behavioral signal's features. A similarity threshold is set; only segments with similarities exceeding this threshold are selected as segments to be emphasized. Through these steps, the computer system can identify multiple segments that fit the target behavioral signal and require emphasis.

[0109] In step S142, after identifying the signal segments to be emphasized, the computer system assesses the continuity and tightness between these signal segments. The continuity assessment parameter is a quantitative indicator used to measure whether the signal segments are closely connected without significant discontinuities or jumps. This is crucial for understanding the continuity and stability of patient behavior.

[0110] The continuity assessment parameter can be calculated through the following steps: For each signal segment to be emphasized, the computer system records its start and end coordinates (e.g., timestamps) within the selected category of behavioral signals. For any two adjacent signal segments to be emphasized, the time interval (or spatial distance, if the signal is multidimensional) between them is calculated. One or more reference coordinates are selected, which can represent the importance of the signal segment or the focus of the analysis. For example, in a heart rate signal, points of abrupt heart rate changes can be selected as reference coordinates. The continuity assessment parameter is then calculated, for example, by first calculating the weighted sum of the intervals between all adjacent signal segments, with weights set according to the importance of the signal segment or the analytical requirements. For example, a signal segment containing critical information (such as a sudden change in heart rate) can be given a higher weight. The influence of the reference coordinates on the continuity of the signal segment to be emphasized is considered. If the signal segment to be emphasized contains reference coordinates, or is very close to reference coordinates, the value of the continuity assessment parameter can be increased. The weighted sum of the signal segment intervals is combined with the influence of the reference coordinates to obtain the final continuity assessment parameter. For example, the following formula can be used:

[0111] ;

[0112] in, It is the weight of the spacing between the i-th signal segments. It is the spacing between the i-th and i+1-th signal segments, and m is the number of reference coordinates. It is the effect of the j-th reference coordinate on continuity.

[0113] For example, continuing with the heart rate signal, suppose three signal segments to be emphasized have been identified: A, B, and C, with their start and end coordinates as follows:

[0114] Signal segment A: Signal segment B: Signal segment C: .

[0115] in, Suppose that signal segment B contains a significant heart rate variability point, and therefore it is given higher weight.

[0116] The steps for calculating the continuity assessment parameters are as follows:

[0117] 1. Calculate the signal segment spacing:

[0118] 2. Determine the influence of reference coordinates: Since signal segment B contains a point of sudden heart rate change, additional reference coordinate influence values ​​are given.

[0119] 3. Calculate continuity assessment parameters: ;

[0120] in, These are the spacing weights between signal segments A and B, and between signal segments B and C, respectively. It is the effect of the sudden heart rate change point contained in signal segment B on continuity.

[0121] In step S143, after determining the continuity evaluation parameters, the computer system compares them with preset standards. If the continuity evaluation parameters meet or exceed the preset standards, it indicates that there is a high degree of continuity and tightness among these signal segments to be emphasized, and therefore it is worth emphasizing them.

[0122] For example, continuing with the heart rate signal, assume the preset continuity assessment parameter standard is 10. If the continuity assessment parameter calculated through the above steps is greater than or equal to 10, the computer system will emphasize signal segments A, B, and C. Emphasis can be placed on changing visual attributes such as the signal segments' color, thickness, and flashing frequency, or adding annotations or labels to the chart to make these signal segments more easily noticeable to the user.

[0123] The purpose of this presentation is not only to highlight signal segments that match the detected behavioral signals, but more importantly, to help users understand the overall patterns and trends of patient behavior by demonstrating the continuity and coherence between these signal segments. This is especially important for healthcare professionals, as they need to develop personalized treatment plans and interventions based on patients' behavioral patterns.

[0124] In one implementation, step S120 involves obtaining the matching degree between the target patient's behavioral signals and the behavioral signals to be detected, based on the influence factors of different categories of behavioral signals, including:

[0125] Step S121: Decompose the signal segment to be detected into a signal segment;

[0126] Step S122: Obtain the basic matching degree between different categories of behavioral signals and the behavioral signals to be detected based on the signal segments to be detected;

[0127] Step S123: Based on the basic matching degree between different categories of behavioral signals and the behavioral signals to be detected, and the influence factors of different categories of behavioral signals, obtain the matching degree between different categories of behavioral signals and the behavioral signals to be detected.

[0128] In step S121, the computer system breaks down the behavior signal to be detected into smaller signal segments to facilitate subsequent calculation of the basic matching degree. Signal segmentation can be based on a time window, the number of data points, or other custom rules. Each segment represents a subsequence of the behavior signal to be detected, used for matching analysis with behavior signals of different categories.

[0129] For example, suppose the behavioral signal to be detected is a continuous heart rate data segment spanning 24 hours. To segment the signal, the computer system can set a fixed-length time window (e.g., 1 hour) and divide the heart rate data according to this time window, resulting in multiple 1-hour heart rate signal segments. Each signal segment contains heart rate data points within that time period.

[0130] In step S122, after the signal segmentation is completed, the computer system calculates the basic matching degree between each signal segment to be detected and different categories of behavioral signals. The basic matching degree measures the similarity or correlation between signal segments and can be calculated by various methods, such as cosine similarity, Euclidean distance, and dynamic time warping (DTW).

[0131] For example, continuing with heart rate data, suppose there are two categories of behavioral signals: blood pressure signals and step count signals. For each heart rate signal segment to be detected, the computer system can calculate its baseline matching degree with the blood pressure signal segment and step count signal segment within the corresponding time period.

[0132] Calculate cosine similarity: Assume that the heart rate signal segment, blood pressure signal segment, and step count signal segment can all be represented as vectors, such as the heart rate signal segment vector. Blood pressure signal segment vector Step number signal segment vector Cosine similarity can be calculated using the following formula:

[0133] ;

[0134] ;

[0135] Where · represents the vector dot product, and |·| represents the vector magnitude.

[0136] In some cases, certain parts of a signal segment may be more important than others. For example, in a heart rate signal, outliers (such as heart rate variability points) may be more valuable for analysis than normal values. Therefore, when calculating baseline fit, a computer system can assign higher weights to these important parts. This can be achieved by adjusting the element values ​​of the signal segment vector or by introducing additional weight vectors.

[0137] In step S123, after obtaining the basic matching degree, the computer system combines the influence factors of different categories of behavioral signals to calculate the final matching degree. The influence factors reflect the importance of each category of behavioral signals in a specific analysis scenario and are usually determined based on expert knowledge, historical data, or machine learning models.

[0138] Influence factors can be determined in several ways. First, domain experts can score different categories of behavioral signals based on their experience, with higher scores indicating greater importance. Second, by analyzing historical data, a computer system can identify the category of behavioral signals most relevant to the signal to be detected and assign it a higher influence factor. Third, a classification or regression model can be trained, using the behavioral signal category and the signal to be detected as input, and actual analysis results or expert evaluations as output, to learn the influence factor for each category of behavioral signal.

[0139] Once the influencing factors are determined, the computer system can calculate the final matching degree using the following formula: .in, This represents the category of the i-th behavior signal. This indicates the basic matching degree between the signal of this category and the signal of the behavior to be detected. This represents the influence factor of this type of signal.

[0140] For example, suppose the baseline matching degree between heart rate and blood pressure signals is 0.8, and the baseline matching degree with step count signals is 0.6. Also, suppose that in the heart rate signal analysis scenario, the influence factor for blood pressure signals is 1.2 (because blood pressure and heart rate have a direct physiological correlation), and the influence factor for step count signals is 0.8 (because although step count is related to health, its direct correlation with heart rate is weak). Then the final matching degree calculation is as follows: Blood pressure signal matching degree: 0.8 × 1.2 = 0.96. Step count signal matching degree: 0.6 × 0.8 = 0.48.

[0141] By comparing these final matching scores, the computer system can determine which categories of behavioral signals are most relevant to the behavioral signal to be detected, thus providing support for subsequent analysis and decision-making.

[0142] In one implementation, step S122, obtaining the basic matching degree between different categories of behavioral signals and the behavioral signals to be detected based on the signal segment to be detected, includes:

[0143] Step S1221: Retrieve target signal segments that match the signal segment to be detected from different categories of behavioral signals;

[0144] Step S1222: Based on the priority of the signal segment to be detected and the priority of the target signal segment that matches the signal segment to be detected in different categories of behavioral signals, determine the basic matching degree between the behavioral signals of different categories and the behavioral signals to be detected.

[0145] In step S1221, the computer system's task is to find target signal segments that match the signal segment to be detected in terms of time, content, or features. These target signal segments will be used for subsequent basic matching degree calculations. To achieve this goal, the computer system takes a series of measures to ensure the suitability of the signal segments.

[0146] Fit criteria can be, for example, time synchronization. For time-series data such as heart rate and blood pressure, a computer system can check the degree of temporal overlap between the target signal segment and the signal segment to be detected. If two signal segments have significant temporal overlap and similar trends, they are considered a good match. Alternatively, content similarity can be used. For non-time-series data such as text and images, a computer system can determine their fit by calculating the similarity between signal segments (e.g., cosine similarity, Jaccard similarity, etc.). The higher the similarity, the more likely the signal segments are to be a good match. Another criterion can be feature consistency. For example, in some cases, a computer system can extract features of the signal segments (e.g., mean, variance, peak value, etc.) and determine the fit by comparing the consistency of these features.

[0147] For example, suppose the signal segment to be detected is a continuous heart rate data segment spanning from 9:00 AM to 10:00 AM. The computer system finds a signal segment that matches this heart rate data segment from blood pressure and step count signals.

[0148] For blood pressure signals, the computer system can examine the blood pressure data segment between 9 a.m. and 10 a.m. If this blood pressure data segment significantly overlaps with the heart rate data in time, and the trend of blood pressure change is similar to that of heart rate change (e.g., blood pressure also increases when heart rate increases), then this blood pressure data segment can be considered as the target signal segment for matching.

[0149] For step count signals, computer systems can calculate the similarity between step count signals and heart rate signals. For example, the Dynamic Time Warping (DTW) algorithm can be used to compare the shape similarity between a segment of step count signal and a segment of heart rate signal. If the similarity is high, this segment of step count data can also be considered as the target signal segment for adaptation.

[0150] In step S1222, after finding the suitable target signal segment, the next step of the computer system is to determine the basic matching degree based on the priority of the signal segment. The priority reflects the importance or criticality of the signal segment in the analysis and can be determined based on the content, characteristics, or analysis results of the signal segment.

[0151] Prioritization can be based on several methods. For example, it can be based on signal segment content. Some signal segments may contain crucial information such as outliers, peaks, or troughs, which are essential for the analysis. Therefore, signal segments containing this crucial information can be assigned higher priority. Alternatively, it can be based on signal segment characteristics. Certain characteristics of a signal segment (such as mean, variance, and kurtosis) may be more important for the analysis. For instance, in heart rate analysis, an abnormally high mean heart rate may be more valuable for analysis than a normal mean heart rate. Another method is based on the analysis results. In some cases, the computer system may have already performed a preliminary analysis of the signal segments and drawn certain conclusions or findings. These conclusions or findings can guide the priority determination. For example, if a signal segment is analyzed to be highly correlated with a certain disease state, it should be assigned higher priority.

[0152] Once the priority of the signal segments is determined, the computer system can use the following formula to calculate the basic matching degree: ;in, Let represent the i-th behavioral signal category, and n represent the number of matching target signal segments found in that category. This represents the priority weight of the j-th adapted target signal segment. This represents the j-th adapted target signal segment. Indicates the signal segment to be detected. This represents a function that calculates the similarity between two signal segments.

[0153] For example, continuing with heart rate data, suppose we have already found target signal segments in the blood pressure and step count signals that match the heart rate signal segment to be detected. Now, we need to calculate the baseline matching degree based on the priority of these signal segments.

[0154] Assume that blood pressure signal segment 1 has a priority weight of 0.8 (because it contains a significant blood pressure peak), and blood pressure signal segment 2 has a priority weight of 0.6 (because it only represents ordinary blood pressure fluctuations). Assume that step count signal segment 1 has a priority weight of 0.5 (because its similarity to the heart rate signal segment is not particularly high), and step count signal segment 2 has a priority weight of 0.7 (because it is highly synchronized with the heart rate signal segment in time). Assume that the similarity between each target signal segment and the heart rate signal segment to be detected has already been calculated.

[0155] The above formula can now be used to calculate the baseline matching degree between blood pressure and step count signals and the heart rate signal to be detected: .

[0156] By comparing these two basic matching degrees, the computer system can determine which category of behavioral signal is more relevant to the behavioral signal to be detected, thereby providing support for subsequent analysis and decision-making.

[0157] In summary, steps S1221 and S1222 together constitute the process of calculating the basic matching degree between different categories of behavioral signals and the behavioral signal to be detected. Through fit judgment and priority determination, the computer system can accurately assess the similarity and importance between signal segments, providing strong data support for subsequent analysis.

[0158] In one implementation scheme, if there are multiple target patient behavioral signals, then the method also includes:

[0159] Step S150: Based on the adaptation mechanism between the behavioral signal to be detected and the behavioral signal of each target patient, determine the influencing factor of each target patient's behavioral signal;

[0160] Step S160: Based on the influencing factors of each target patient's behavioral signal and the matching degree between the different categories of behavioral signals covered by each target patient's behavioral signal and the behavioral signal to be detected, obtain the matching degree between each target patient's behavioral signal and the behavioral signal to be detected;

[0161] Step S170: Based on the matching degree between each target patient's behavioral signal and the behavioral signal to be detected, list the multiple target patient behavioral signals in sequence.

[0162] In step S150, the computer system determines the influence factor for each target patient's behavioral signal. The influence factor is a quantitative indicator used to measure the importance or weight of the target patient's behavioral signal in the overall analysis. To accurately determine the influence factor, the computer system considers the adaptation mechanism between the behavioral signal to be detected and each target patient's behavioral signal.

[0163] The matching mechanism refers to the way or rule that establishes a connection between the behavioral signal to be detected and the behavioral signal of the target patient. These mechanisms can be based on a variety of factors, including but not limited to temporal synchronization, content similarity, and feature consistency. For example, in terms of temporal synchronization, if two signals have significant temporal overlap and similar trends, they are considered to be matched. In terms of content similarity, if two signals contain the same or similar information (such as keywords, patterns, etc.), they are also considered to be matched.

[0164] Computer systems can assign different weights to each target patient behavior signal based on the strength and complexity of the adaptation mechanism. For example, if the adaptation mechanism between a target patient behavior signal and the behavior signal to be detected is very clear and strong (such as direct time synchronization and content matching), then a higher weight can be assigned to that signal.

[0165] In some cases, computer systems can use machine learning models to predict the influencing factors of each target patient's behavioral signal. These models can be trained on historical data and learn how to map adaptation mechanisms to influencing factors. For example, regression or classification models can be used to predict influencing factors, where input features may include multiple aspects of the adaptation mechanism (such as time synchronization, content similarity, feature consistency, etc.), and the output is the influencing factor of the target patient's behavioral signal.

[0166] In step S160, the computer system combines the matching degree between the different categories of behavioral signals and the behavioral signals to be detected calculated in step S120, as well as the influence factors determined in step S150, to calculate the overall matching degree between the behavioral signals of each target patient and the behavioral signals to be detected.

[0167] The overall match rate can be calculated using the following formula:

[0168] ;in, Let represent the behavioral signal of the i-th target patient, and n represent the number of different categories of behavioral signals covered in this signal. This represents the influence factor of the j-th category of behavioral signals in the i-th target patient's behavioral signals. This represents the matching degree between the j-th category behavior signal and the behavior signal to be detected (calculated in step S120).

[0169] In step S170, the computer system sorts the multiple target patient behavioral signals according to the overall matching degree calculated in step S160 and lists them in order. This helps medical professionals quickly identify the target patient behavioral signals most relevant to the behavioral signals to be detected, thereby enabling more efficient subsequent analysis.

[0170] The computer system can sort the overall matching scores from high to low, placing the behavioral signals of the target patient with the highest matching scores at the top. This allows healthcare professionals to focus first on signals most relevant to the behavioral signals being detected.

[0171] In some cases, the computer system can also set a matching threshold. Only the behavioral signals of target patients whose overall matching score exceeds this threshold will be listed. This helps reduce noise interference and improve analysis efficiency.

[0172] For example, suppose there are three target patient behavioral signals (patient A, patient B, and patient C), each signal encompassing two categories of behavioral signals (such as heart rate and blood pressure). In step S120, the matching degree between each category of behavioral signal and the target behavioral signal has been calculated. In step S150, an influence factor is assigned to each category of behavioral signal within each target patient behavioral signal according to an adaptation mechanism. Now, the overall matching degree between each target patient behavioral signal and the target behavioral signal is calculated and ranked:

[0173] For patient A:

[0174] Heart rate signal matching degree: 0.8; Blood pressure signal matching degree: 0.6; Heart rate signal influence factor: 0.7 (based on adaptation mechanism allocation); Blood pressure signal influence factor: 0.5 (based on adaptation mechanism allocation); Overall matching degree (patient A): 0.7×0.8 +0.5×0.6=0.86.

[0175] For patient B:

[0176] Heart rate signal matching degree: 0.7; Blood pressure signal matching degree: 0.5; Heart rate signal influence factor: 0.8 (based on adaptation mechanism allocation); Blood pressure signal influence factor: 0.4 (based on adaptation mechanism allocation); Overall matching degree (patient B): 0.8×0.7 +0.4×0.5=0.76.

[0177] For patient C:

[0178] Heart rate signal matching degree: 0.6; Blood pressure signal matching degree: 0.4; Heart rate signal influence factor: 0.6 (based on adaptation mechanism allocation); Blood pressure signal influence factor: 0.3 (based on adaptation mechanism allocation); Overall matching degree (patient C): 0.6×0.6 +0.3×0.4=0.48.

[0179] The target patient's behavioral signals are listed below in descending order of overall match:

[0180] 1. Patient A (overall match: 0.86); 2. Patient B (overall match: 0.76); 3. Patient C (overall match: 0.48).

[0181] Through the above steps, the signal most relevant to the behavioral signal to be detected can be accurately identified from multiple target patient behavioral signals, providing orderly data support for medical workers.

[0182] In one implementation, step S110, acquiring the target patient's behavioral signal that matches the behavioral signal to be detected, includes:

[0183] Step S111: Decompose the signal segment of the behavior signal to be detected to obtain a set of signal segments to be detected;

[0184] Step S112: Divide the signal segments in the set of signal segments to be detected into clusters to obtain multiple signal segment clusters;

[0185] Step S113: Obtain the target patient behavior signals that are adapted to multiple signal segment clusters respectively, with one signal segment cluster corresponding to one adaptation mechanism.

[0186] In step S111, the computer system splits the signal to be detected into multiple smaller signal segments to form a set of signal segments to be detected. Signal segment splitting is one of the key steps in data preprocessing, which helps the computer system to analyze the characteristics and patterns of the signal to be detected in greater detail.

[0187] Signal segmentation methods include time-window-based segmentation. The computer system can set a fixed-length time window (such as 1 minute, 5 minutes, etc.) and then segment the behavioral signal to be detected according to this time window. For example, if the behavioral signal to be detected is a continuous heart rate data spanning 24 hours, the computer system can segment it into 288 10-minute signal segments (assuming there are 6 10-minute signal segments per hour).

[0188] Alternatively, based on event-triggered segmentation, in some cases, the behavioral signal to be detected may contain specific event markers (such as outliers, peaks, etc.). The computer system can then segment the signal into different segments based on these event markers. For example, in heart rate data, the computer system can segment the signal into a new segment whenever the heart rate exceeds a certain threshold.

[0189] Alternatively, in addition to time windows, computer systems can also segment signals based on the number of data points. For example, each signal segment can be set to contain 100 data points, and the signal to be detected can be segmented according to this number.

[0190] For example, suppose the behavioral signal to be detected is a continuous heart rate data segment spanning one hour. The computer system sets a 5-minute time window and divides the heart rate data into 12 signal segments. Each signal segment contains heart rate data points within that time period, forming a set of signal segments to be detected.

[0191] In step S112, after obtaining the set of signal segments to be detected, the next step of the computer system is to cluster these signal segments to form multiple signal segment clusters. The purpose of clustering is to group signal segments with similar characteristics or patterns into one category, so that the target patient behavioral signals that match the behavioral signals to be detected can be identified more accurately in the future.

[0192] Clustering can be achieved through features-based clustering. The computer system can extract features from each signal segment (such as mean, variance, kurtosis, skewness, etc.) and then use clustering algorithms (such as K-means, DBSCAN, etc.) to group these feature vectors into clusters. Signal segments with similar features will be grouped together.

[0193] Alternatively, pattern-based clustering can be used. In addition to features, computer systems can also consider pattern similarity between signal segments. For example, the Dynamic Time Warping (DTW) algorithm can be used to compare the shape similarity between signal segments and group signal segments with similar shapes into a single cluster.

[0194] Alternatively, clustering can be based on expert knowledge. In some cases, domain experts may be able to divide signal segments into different clusters based on experience. The computer system can accept expert input and perform clustering accordingly. For example, continuing with heart rate data, the computer system extracts the mean and variance of each signal segment as features, and then uses the K-means clustering algorithm to cluster these feature vectors. Suppose that three clusters are ultimately obtained, representing low heart rate, normal heart rate, and high heart rate signal segments, respectively.

[0195] In step S113, after obtaining multiple signal segment clusters, the final step of the computer system is to acquire the target patient behavior signal that is adapted to each signal segment cluster. The adaptation mechanism refers to the way or rule that establishes a connection between the signal segment cluster and the target patient behavior signal. Different signal segment clusters may correspond to different adaptation mechanisms.

[0196] For example, time synchronization can be used; that is, for time series data such as heart rate and blood pressure, the time synchronization between signal clusters and the target patient's behavioral signals is an important adaptation mechanism. If two signals have significant temporal overlap and similar trends, they are, for example, a good match.

[0197] Alternatively, content similarity can be used. In some cases, the content similarity between signal clusters and target patient behavioral signals may be more important. For example, in text analysis, if two signal clusters contain the same keywords or phrases, they may be matched with the same target patient behavioral signals.

[0198] Alternatively, feature consistency can be employed; the consistency of features between signal segment clusters and target patient behavior signals is another possible adaptation mechanism. For example, in image analysis, if two signal segment clusters have similar texture, color, or other features, they may be adapted to the same target image signal. Continuing with the heart rate data example, for signal segment clusters representing low heart rates, the computer system can search for target patient behavior signals that also contain low heart rate signal segments as adaptation signals. For signal segment clusters representing normal heart rates, it can search for target patient behavior signals that contain normal heart rate signal segments. And for signal segment clusters representing high heart rates, it can search for target patient behavior signals that contain abnormal heart rate signal segments (such as sudden increases or decreases in heart rate). These adaptation mechanisms may be determined based on factors such as time synchronization, content similarity, or feature consistency.

[0199] Through the above steps, the computer system can accurately identify the signal segment most relevant to the target behavioral signal from a large number of target patient behavioral signals, and provide strong support for subsequent data analysis and pattern recognition.

[0200] In one implementation, step S111 involves segmenting the signal to be detected behavior signal into signal segments to obtain a set of signal segments to be detected, including:

[0201] Step S1111: Classify the behavior patterns of the behavior signals to be detected to obtain behavior pattern labels for the behavior signals to be detected;

[0202] Step S1112: Segment the behavior signal to be detected, and generate a set of signal segments to be detected based on the behavior pattern label and the signal segments obtained from the signal segmentation.

[0203] In the implementation of step S111, the goal of the computer system is to segment the signal to be detected behavior signal into signal segments to form a set of signal segments to be detected.

[0204] In step S1111, the computer system performs behavioral pattern classification on the behavioral signals to be detected, so as to assign a behavioral pattern label to each signal segment. The behavioral pattern label is a general description of the characteristics of the signal segment, which helps the computer system better understand the content and properties of the signal segment.

[0205] Behavioral pattern classification can be achieved through rule-based classification, where the computer system categorizes the detected behavioral signals according to preset rules or thresholds. For example, in heart rate signal analysis, a heart rate threshold can be set (e.g., 60 to 100 beats per minute is considered normal, below 60 beats per minute is bradycardia, and above 100 beats per minute is tachycardia). The computer system can then iterate through the entire heart rate signal, classifying it as normal, bradycardia, or tachycardia based on the value of each data point.

[0206] Alternatively, classification can be based on machine learning. Besides rule-based classification, computer systems can use machine learning models to classify behavioral signals to be detected. These models can learn the mapping between signal segments and behavioral patterns through training and automatically classify new signal segments. For example, algorithms such as Support Vector Machines (SVM), Random Forests, or neural networks can be used to train classification models.

[0207] In some cases, the behavioral signal to be detected may lack explicit labels or category information. In such situations, computer systems can use unsupervised learning methods (such as clustering algorithms) to discover similarities between signal segments and classify them based on these similarities. While this method cannot directly obtain behavioral pattern labels, it can help the computer system identify the underlying structure and patterns between signal segments. For example, suppose the behavioral signal to be detected is a continuous heart rate data segment. The computer system first uses a rule-based classification method to divide the signal into three categories based on heart rate values: normal heart rate (60-100 beats / minute), bradycardia (<60 beats / minute), and tachycardia (>100 beats / minute). Then, the computer system assigns a behavioral pattern label to each category, such as "normal," "bradycardia," and "tachycardia."

[0208] In step S1112, after obtaining the behavior pattern label, the next step of the computer system is to segment the behavior signal to be detected and generate a set of signal segments to be detected based on the behavior pattern label and the signal segments. Signal segmentation is the process of cutting continuous signal data into multiple shorter signal segments, which helps the computer system to analyze the characteristics and patterns of the signal more precisely.

[0209] Signal segmentation can be achieved through time-window-based segmentation. The computer system can set a fixed-length time window (such as 1 minute, 5 minutes, etc.) and then divide the signal to be detected according to this time window. The data points within each time window will form a signal segment.

[0210] Alternatively, segmentation can be based on event triggers. In some cases, the signal being detected may contain specific event markers (such as outliers, peaks, etc.). The computer system can then segment the signal into different segments based on these event markers. For example, in heart rate data, a new signal segment can be triggered whenever the heart rate exceeds a certain threshold.

[0211] Alternatively, based on adaptive segmentation, in addition to time-window and event-triggered segmentation methods, computer systems can also use adaptive segmentation to process complex signal data. This method dynamically adjusts the segment length and position according to the characteristics and patterns of the signal to better capture key information in the signal.

[0212] After signal segmentation, the computer system associates each signal segment with its corresponding behavioral pattern label, generating a set of signal segments to be detected. This set contains all signal segments and their corresponding behavioral pattern labels, providing a foundation for subsequent data analysis and pattern recognition. For example, continuing with heart rate data, the computer system sets a 5-minute time window and segments the heart rate data into multiple signal segments. Then, based on the behavioral pattern labels obtained in step S1111, it assigns a label (such as "normal," "bradycardia," or "tachycardia") to each signal segment. Finally, the computer system generates a set of signal segments to be detected, containing all signal segments and their corresponding behavioral pattern labels.

[0213] Through the above steps, the computer system can perform refined segmentation and classification of the behavioral signals to be detected, and generate a set of signal segments containing behavioral pattern labels. This provides strong support for subsequent data analysis and pattern recognition, helping the computer system to more accurately identify key information and patterns in the signals.

[0214] In one implementation, step S130 involves filtering behavioral signals of a selected category from different categories of behavioral signals based on the matching degree between different categories of behavioral signals and the behavioral signal to be detected, including:

[0215] Step S131: Filter out the target category behavior signals and the behavior signals of unselected categories other than the target category from the different categories of behavior signals;

[0216] Step S132: Based on the matching degree between different categories of behavioral signals and the behavioral signals to be detected, filter out behavioral signals of unselected categories from the behavioral signals to obtain behavioral signals of undetermined categories;

[0217] Step S133: Use the behavior signals of the undetermined category and the behavior signals of the target category as the behavior signals of the selected category.

[0218] In the implementation of step S130, the goal of the computer system is to select the signal category that best matches the signal to be detected from different categories of behavioral signals, as the selected category for subsequent analysis.

[0219] In step S131, the computer system performs a preliminary screening of all different categories of behavioral signals to determine which categories are target categories and which are unselected categories. Target categories are usually predetermined based on domain knowledge, expert opinions, or historical data, and they have a high correlation or importance with the behavioral signals to be detected.

[0220] Target categories can be determined in ways such as based on domain knowledge. In some cases, domain experts may be able to directly determine which behavioral signal categories are most relevant to the behavioral signal being detected based on their expertise. For example, in heart rate monitoring, heart rate and blood pressure signals are often the target categories when analyzing heart rate abnormalities.

[0221] Alternatively, based on historical data, computer systems can analyze historical data to determine which behavioral signal categories have a high correlation with the behavioral signal to be detected. For example, association rule mining algorithms (such as the Apriori algorithm) can be used to discover frequent itemsets between behavioral signal categories and the behavioral signal to be detected, thereby determining the target category.

[0222] Alternatively, based on machine learning models, in certain complex scenarios, computer systems can use machine learning models to predict which categories of behavioral signals are most likely related to the behavioral signal to be detected. For example, a classification model can be trained, using the category of the behavioral signal as input features and the presence or absence of the behavioral signal to be detected as the output label, to determine the target category through model prediction. For instance, suppose we are analyzing the electrocardiogram data of a heart patient, and the behavioral signal to be detected is the change in heart rate during a heart attack. Based on domain knowledge and historical data, heart rate and blood pressure signals can be identified as the target categories highly correlated with heart attacks. Other behavioral signals (such as step count, sleep quality, etc.) are considered as unselected categories outside the target categories.

[0223] In step S132, after determining the target category and the unselected categories, the next step of the computer system is to filter out signal categories with a high degree of matching with the behavioral signal to be detected from the behavioral signals of the unselected categories, based on the matching degree calculated in step S120, and designate them as undetermined categories. Undetermined categories refer to those behavioral signal categories that, although not initially identified as target categories, show a high correlation in actual data analysis.

[0224] One method for filtering pending categories is to set a matching threshold. The computer system can set a matching threshold, and only unselected categories with a matching degree exceeding the threshold will be filtered into pending categories. The threshold can be adjusted according to actual needs and data characteristics.

[0225] The computer system can sort the matching scores of all unselected categories and select the top N categories with the highest matching scores as candidate categories (N is a preset integer). This method ensures that the selected candidate categories have high relevance and importance.

[0226] Alternatively, machine learning models can be used for filtering. In some cases, computer systems can employ machine learning models to assist in filtering candidate categories. For example, a regression model can be trained to predict the match between each unselected category and the behavioral signal to be detected, and the candidate categories can be filtered based on the predictions. For instance, let's continue with the example of electrocardiogram (ECG) data from a heart patient. Assume that the match between heart rate signals, blood pressure signals, and other unselected categories (such as steps, sleep quality, etc.) and the heart rate change signal to be detected has been calculated. By setting a match threshold or a ranking selection method, the step count signal can be filtered out from the unselected categories as a candidate category, because the step count signal may have some correlation with heart rate changes during a heart attack (e.g., reduced activity before a heart attack may lead to a decrease in steps).

[0227] In step S133, the computer system merges the behavioral signals of the undetermined category with the behavioral signals of the target category to form the final selected category set. The selected category set contains all behavioral signal categories that are highly correlated with the behavioral signals to be detected, and these will be the focus of subsequent data analysis.

[0228] Category merging can be achieved by directly combining the candidate category and the target category into a new set, which is then designated as the selected category. Alternatively, in some cases, the computer system can assign different weights to different category signals based on their importance or matching degree, and consider these weights during the merging process. For example, a weighted average method can be used to calculate the overall matching degree or importance score of the selected category.

[0229] For example, let's continue using electrocardiogram (ECG) data from a heart disease patient. In step S132, the step count signal has already been selected as a candidate category. Now, the step count signal is merged with the heart rate and blood pressure signals (target categories) to form the final set of selected categories. This set will be the focus of subsequent data analysis to further explore the relationship between heart attacks and heart rate, blood pressure, and step count.

[0230] Through the above steps, the computer system can accurately filter out the signal category that best matches the signal to be detected from different categories of behavioral signals, providing strong support for subsequent data analysis and pattern recognition.

[0231] In one implementation, step S132, based on the matching degree between different categories of behavioral signals and the behavioral signal to be detected, filters out behavioral signals of unselected categories from the behavioral signals to obtain behavioral signals of a pending category, including:

[0232] Step S1321: Based on the matching degree between different categories of behavioral signals and the behavioral signals to be detected, select the original undetermined category behavioral signals whose matching degree is higher than the matching degree threshold from the behavioral signals of unselected categories;

[0233] Step S1322: Take the behavior signals of the original undetermined categories corresponding to the top K matching degrees in descending order as the behavior signals of the undetermined categories, where K≥1.

[0234] In step S1321, the computer system determines a matching threshold and, based on this threshold, filters out signal categories with higher matching degrees from the behavioral signals of unselected categories. These signal categories will be regarded as behavioral signals of the original undetermined categories, and they may be further confirmed as undetermined categories in subsequent steps.

[0235] One method for determining the matching threshold is based on statistical distribution. The computer system can calculate the statistical distribution (such as mean, standard deviation, etc.) of the matching degree between all unselected behavior signals and the behavior signal to be detected, and determine a reasonable matching threshold based on this distribution. For example, the mean plus one or two times the standard deviation can be chosen as the threshold to ensure that the selected signal categories have a certain degree of statistical significance.

[0236] Alternatively, a domain-knowledge-based approach may be used. In some cases, domain experts may directly provide a reasonable matching threshold based on their expertise. This threshold may be based on past research experience, experimental results, or industry standards.

[0237] Alternatively, based on machine learning models, the computer system can use these models to predict a reasonable threshold for matching. This model can be trained on historical data, learning the relationship between matching degree and the importance of the signal category, and outputting a predicted threshold. For example, suppose we are analyzing the blood glucose data of a diabetic patient, and the behavioral signal to be detected is an event where the patient's blood glucose level rose abnormally. There is a series of behavioral signals without selected categories, including diet records, exercise records, sleep records, etc. First, the computer system calculates the matching degree between these signals and the behavioral signal to be detected. Then, based on a statistical distribution method, the mean plus one standard deviation is selected as the matching degree threshold. In this way, all signal categories with matching degrees higher than this threshold (such as diet records) are filtered out as behavioral signals in the initial undetermined category.

[0238] In step S1322, after determining the initial candidate categories of behavioral signals, the next step of the computer system is to sort these signal categories according to their matching degree and select the top K signal categories with the highest matching degree as the final candidate categories. This process helps ensure that the selected candidate categories have the highest correlation with the behavioral signals to be detected.

[0239] The computer system sorts all the behavioral signals initially categorized into undetermined classes based on their matching degree, from highest to lowest. The signal categories with the highest matching degree are then listed first. After sorting, the computer system selects the top K signal categories with the highest matching degree as undetermined classes. The value of K can be determined based on actual needs and data characteristics. For example, if the dataset contains many signal categories, but only a few of the most important categories are desired for analysis, then K can be set relatively small; conversely, if a more comprehensive understanding of the data is desired, then K can be set relatively large.

[0240] For example, let's continue using blood glucose data from diabetic patients as an example. In step S1321, dietary records have already been selected as behavioral signals for the initial undetermined category. Assume that other unselected behavioral signals (such as exercise records, sleep records, etc.) also reach the matching threshold and are selected as behavioral signals for the initial undetermined category. Now, the computer system sorts these signal categories according to their matching degree from high to low. Assume the sorting result is: dietary records (matching degree 0.9), exercise records (matching degree 0.7), and sleep records (matching degree 0.5). If K=2 is selected, then dietary records and exercise records will be selected as the final undetermined categories because they have the highest matching degree with the behavioral signal to be detected.

[0241] Through the above steps, the computer system can filter out signal categories with a high degree of matching with the behavioral signal to be detected from the behavioral signals that have not been selected as categories, and use these categories as behavioral signals to be determined. This process not only considers the correlation between the signal category and the behavioral signal to be detected, but also ensures that the selected categories to be determined have the highest importance through sorting and selection.

[0242] In one implementation, step S140, which emphasizes and displays the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected, further includes:

[0243] Step S144: Select target category signal segments from the set of signal segments to be detected. The set of signal segments to be detected includes signal segments obtained by segmenting the signal segments of the behavior to be detected.

[0244] Step S145: Emphasize and display the signal segments in the selected category of behavioral signals that match the target category signal segments.

[0245] In step S144, the computer system filters out target category signal segments from the set of signal segments to be detected. The set of signal segments to be detected is obtained in step S111 by segmenting the signal segments of the behavior signal to be detected. It contains multiple signal segments, each representing a subsequence of the behavior signal to be detected. Target category signal segments refer to those signal segments that are closely related to a specific analysis target or point of interest.

[0246] Computer systems can extract features of each signal segment (such as mean, variance, kurtosis, skewness, etc.) and filter signal segments into target categories based on these features. For example, in heart rate monitoring, if the analysis goal is to identify abnormal heart rate events, the computer system can filter out signal segments with heart rate values ​​outside the normal range (such as less than 60 beats per minute or more than 100 beats per minute) as target category signal segments.

[0247] In addition to feature-based filtering methods, computer systems can also use pattern matching algorithms to identify target category signal segments. For example, the Dynamic Time Warping (DTW) algorithm can be used to compare the shape similarity between signal segments and filter out signal segments that have a high degree of matching with a preset pattern or template as target category signal segments.

[0248] In some cases, domain experts can directly determine which signal segments belong to the target category based on their expertise. The computer system can then accept expert input and perform the filtering accordingly.

[0249] For example, suppose we are analyzing the electrocardiogram (ECG) data of a heart patient, and the behavioral signal to be detected is the heart rate change during a heart attack. In step S111, the heart rate data has been divided into multiple signal segments, forming a set of signal segments to be detected. Now, the goal of the analysis is to identify abnormal heart rate events during a heart attack. Based on this goal, a feature-based screening method can be used to filter out signal segments with heart rate values ​​outside the normal range (e.g., below 60 beats per minute or above 120 beats per minute) as the target category signal segments.

[0250] In step S145, after filtering out the target category signal segments, the computer system finds signal segments in the selected category of behavioral signals that match these signal segments and highlights them. Fit is typically evaluated based on the similarity or correlation between signal segments, which can be achieved through various methods. The computer system can calculate the similarity (e.g., cosine similarity, Euclidean distance, etc.) between signal segments in the selected category of behavioral signals and signal segments in the target category, and evaluate fit based on this similarity. Higher similarity indicates a better fit between the two signal segments. For time-series data, such as heart rate and blood pressure, the computer system can check the temporal overlap between signal segments in the selected category of behavioral signals and signal segments in the target category. If two signal segments have significant temporal overlap and similar trends, they are, for example, a good match. In addition to similarity and temporal synchronization, the computer system can also extract features of the signal segments (e.g., mean, variance, etc.) and compare the consistency of these features to evaluate fit. Higher feature consistency indicates a greater likelihood of a good match between the two signal segments.

[0251] Computer systems can emphasize and display appropriate signal segments by changing their visual attributes, such as color, thickness, and flashing frequency. For example, different colors can be used to distinguish different categories of signal segments, or bold lines can be used to highlight appropriate signal segments.

[0252] In addition to visual emphasis, computer systems can also provide interactive display functions, allowing users to view detailed information about the adapted signal segment through operations such as zooming and panning. For example, interactive chart libraries (such as D3.js, ECharts, etc.) can be used to create dynamic, interactive signal display interfaces.

[0253] To further help users understand the meaning and importance of the adapted signal segments, the computer system can also generate corresponding text descriptions and explanations. These descriptions and explanations can be written based on the characteristics, patterns, or analysis results of the signal segments and presented on the user interface.

[0254] For example, let's continue using electrocardiogram (ECG) data from a heart attack patient. In step S144, target category signal segments have been selected, namely signal segments representing abnormal heart rate events during a heart attack. Now, it's necessary to find signal segments within the selected category of behavioral signals (such as heart rate signals, blood pressure signals, etc.) that match these target category signal segments and highlight them. Assuming the selected category of heart rate signals contains a signal segment that matches the target category (such as a sudden increase in heart rate), the computer system can highlight it by changing the color of that signal segment (e.g., using red) or by making it bold. Simultaneously, corresponding text descriptions and explanations can be added to the user interface, such as "Sudden increase in heart rate event, possibly related to a heart attack."

[0255] In one implementation, step S140, which emphasizes and displays the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected, further includes:

[0256] Step S146: Determine the unselected category signal segments in the set of signal segments to be detected, excluding the target category signal segments;

[0257] Step S147: Retrieve signal segments that match the signal segments of unselected categories from the behavioral signals of the selected categories, and obtain a set of candidate signal segments;

[0258] Step S148: Select G signal segments from the candidate signal segment set for emphasis display, where G≥1.

[0259] In step S146, the computer system identifies signal segments from the set of signal segments to be detected that do not belong to the target category. These signal segments are called unselected category signal segments. Unselected category signal segments may contain information related to the signal of the behavior to be detected but not explicitly classified into the target category, and therefore they also have potential analytical value.

[0260] If each signal segment in the set of signal segments to be detected has been assigned a corresponding label (such as "target category", "non-target category", etc.), the computer system can easily identify signal segments without a selected category by checking the label.

[0261] If signal segments lack explicit labels, computer systems can distinguish whether they belong to the target category by analyzing their characteristics (such as mean, variance, kurtosis, skewness, etc.). Signal segments whose characteristics significantly differ from those of the target category are considered unselected category segments. In some cases, computer systems can use machine learning models to predict whether each signal segment belongs to the target category. This model can be trained on historical data and learn how to distinguish between target category and unselected category signal segments.

[0262] For example, suppose we are analyzing the blood glucose data of a diabetic patient. The target category signal segments are those with abnormally high or low blood glucose levels. In the set of signal segments to be detected, besides the target category signal segments, there may be other signal segments with smaller fluctuations in blood glucose levels; these are considered unselected category signal segments.

[0263] In step S147, after identifying the signal segments of the unselected categories, the computer system retrieves signal segments from the behavioral signals of the selected categories that match these signal segments. The fit is typically evaluated based on the similarity or correlation between signal segments, resulting in a set of candidate signal segments that may contain similar information or patterns to the unselected category signal segments.

[0264] The computer system can calculate the similarity (such as cosine similarity, Euclidean distance, etc.) between each signal segment in the selected category of behavioral signals and the signal segments of the unselected category, and select the signal segments with similarity exceeding a certain threshold as candidate signal segments.

[0265] For time-series data, such as heart rate and blood pressure, computer systems can examine the temporal overlap between signal segments in selected categories and those in unselected categories. If two signal segments significantly overlap temporally and exhibit similar trends, they are considered a good match. In addition to similarity and temporal synchronization, computer systems can also extract features of the signal segments (such as mean and variance) and compare the consistency of these features to assess fit. Higher feature consistency indicates a greater likelihood of a good match between the two signal segments.

[0266] For example, let's continue using blood glucose data from diabetic patients as an example. Assume the selected behavioral signals include diet records, exercise records, etc. For each unselected category signal segment (i.e., a segment with relatively small fluctuations in blood glucose levels), the computer system can search the diet and exercise records for signal segments that are time-synchronized and have similar characteristics to these signal segments as candidate signal segments. For example, if a patient's diet and exercise habits are relatively stable within the time period corresponding to a certain unselected category signal segment, the computer system can find similar signal segments in these records as candidate signal segments.

[0267] In step S148, after obtaining the candidate signal segment set, the final step of the computer system is to select G signal segments for emphasis display. The value of G can be determined according to actual needs and analysis objectives; it represents the number of candidate signal segments the user wants to see. The selection process may involve comprehensive consideration of various factors, such as the similarity of signal segments, time synchronization, and feature consistency.

[0268] The computer system can sort candidate signal segments based on indicators such as similarity, time synchronization, or feature consistency, and select the top G signal segments for emphasis display.

[0269] In some cases, users may want to filter candidate signal segments based on their analytical goals and interests. Computer systems can provide interactive interfaces that allow users to customize filtering criteria and display methods. In some complex scenarios, computer systems can use machine learning models to predict which candidate signal segments are most likely to be useful to the user and filter and emphasize the displayed signal segments based on the prediction results.

[0270] For example, let's continue using blood glucose data from diabetic patients. Suppose we set G=3, meaning we want to highlight three candidate signal segments. The computer system can sort these segments based on their similarity and temporal synchronization, and then select the top three segments for emphasis. These segments likely represent the most relevant and important dietary or exercise records to the unselected segments, helping users gain a deeper understanding of the potential causes and influencing factors of blood glucose fluctuations.

[0271] In one implementation, step S141 involves retrieving multiple signal segments to be emphasized from the selected category of behavioral signals, each segment being adapted to the behavioral signal to be detected.

[0272] Step S1411: Obtain the first signal segment and the second signal segment whose signal coordinates are adjacent in the signal to be detected;

[0273] Step S1412: Retrieve the third signal segment that matches the first signal segment and the fourth signal segment that matches the second signal segment from the selected category of behavioral signals.

[0274] Step S142, determining the continuity evaluation parameters between multiple signal segments to be emphasized by using the signal coordinates of multiple signal segments to be emphasized in the behavioral signals of the selected category, including:

[0275] Step S1421: Determine the continuity evaluation parameters between the third signal segment and the fourth signal segment based on the signal coordinates of the third signal segment in the selected category of behavioral signals and the signal coordinates of the fourth signal segment in the selected category of behavioral signals.

[0276] Step S143: When the continuity evaluation parameters between multiple signal segments to be emphasized reach the preset standard, the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are emphasized and displayed, including:

[0277] Step S1431: When the continuity evaluation parameter between the third signal segment and the fourth signal segment reaches the preset standard, select signal segments that match the behavior signal to be detected from the selected category of behavior signals, and highlight the selected signal segments.

[0278] In step S1411, the computer system finds two signal segments with adjacent signal coordinates in the behavior signal to be detected, referred to as the first signal segment and the second signal segment, respectively. These two signal segments should be closely connected so that corresponding signal segments can be found in the selected category of behavior signals.

[0279] If the behavioral signal to be detected is time-series data (such as heart rate, blood pressure, etc.), the computer system can set a fixed-length time window to divide the signal into multiple signal segments. Then, signal segments within two adjacent time windows are selected as the first and second signal segments. In some cases, the behavioral signal to be detected may contain specific event markers (such as outliers, peaks, etc.). The computer system can divide the signal into different segments based on these event markers and select adjacent event segments as the first and second signal segments. If the characteristics of the behavioral signal to be detected (such as mean, variance, etc.) change significantly within a short period of time, the computer system can divide the signal into different segments based on these points of change and select adjacent characteristic change segments as the first and second signal segments.

[0280] For example, suppose the behavioral signal to be detected is a continuous heart rate data segment spanning one hour. The computer system sets a 5-minute time window, dividing the heart rate data into 12 signal segments. Now, two adjacent signal segments (such as the 5th and 6th signal segments) are selected as the first and second signal segments.

[0281] In step S1412, after determining the first and second signal segments, the next step of the computer system is to retrieve signal segments from the selected category of behavioral signals that are respectively adapted to these two signal segments, referred to as the third and fourth signal segments. Adaptability is typically evaluated based on the similarity or correlation between the signal segments.

[0282] The computer system can calculate the similarity (such as cosine similarity, Euclidean distance, etc.) between each signal segment in the selected category of behavioral signals and the first and second signal segments, and select signal segments with similarity exceeding a certain threshold as the third and fourth signal segments.

[0283] For time series data, a computer system can examine the degree of temporal overlap between a signal segment in a selected category of behavioral signals and a first and second signal segment. If the two signal segments have significant temporal overlap and similar trends, they are considered a good match.

[0284] In addition to similarity and time synchronization, computer systems can also extract features of signal segments (such as mean and variance) and compare the consistency of these features to assess fit. The higher the feature consistency, the more likely the two signal segments are to fit.

[0285] For example, let's continue with heart rate data. Suppose the selected categories of behavioral signals include blood pressure signals and step count signals. For the first signal segment (the 5th heart rate signal segment), the computer system might find a similar time period in the blood pressure signal as the third signal segment; for the second signal segment (the 6th heart rate signal segment), the computer system might find a similar time period in the step count signal as the fourth signal segment.

[0286] In step S1421, the computer system evaluates the continuity between the third and fourth signal segments to determine whether they are closely connected, thus forming a meaningful sequence of signal segments. The continuity evaluation parameter is a quantitative indicator used to measure the closeness and continuity between signal segments.

[0287] For time series data, a computer system can calculate the time interval between the third and fourth signal segments. A shorter time interval indicates better continuity between the two segments. In addition to the time interval, the computer system can also consider the reference value of the signal coordinates. For example, if both the third and fourth signal segments contain important feature changes or event markers, then their continuity may be more important. The computer system can combine the time interval and signal coordinate reference value, calculating continuity evaluation parameters through weighted summation or multiplication. The weights can be determined based on actual needs and analytical objectives.

[0288] For example, let's continue with heart rate data. Suppose the third signal segment is a time interval within the blood pressure signal, and the fourth signal segment is a time interval within the step count signal. The computer system can calculate the time interval between these two time segments and consider whether they both contain significant characteristic changes (such as a sudden rise in blood pressure, a sudden decrease in steps, etc.). Then, based on the time interval and signal coordinate reference, continuity assessment parameters are calculated.

[0289] In step S1431, the computer system determines whether the continuity evaluation parameters between the third and fourth signal segments meet the preset standard. If the standard is met, the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are highlighted and displayed; otherwise, further screening or adjustment of the signal segments may be necessary.

[0290] Computer systems can emphasize and display appropriate signal segments by changing their visual attributes, such as color, thickness, and flashing frequency. For example, different colors can be used to distinguish different categories of signal segments, or bold lines can be used to highlight appropriate signal segments.

[0291] In addition to visual emphasis, computer systems can also provide interactive display functions, allowing users to view detailed information about the adapted signal segment through operations such as zooming and panning. For example, interactive chart libraries (such as D3.js, ECharts, etc.) can be used to create dynamic, interactive signal display interfaces.

[0292] To further help users understand the meaning and importance of the adapted signal segments, the computer system can also generate corresponding text descriptions and explanations. These descriptions and explanations can be written based on the characteristics, patterns, or analysis results of the signal segments and presented on the user interface.

[0293] For example, continuing with heart rate data, suppose the continuity assessment parameters between the third and fourth signal segments meet a preset standard, indicating that these two signal segments are closely connected and form a meaningful signal segment sequence. In this case, the computer system can highlight the signal segments (such as the third and fourth signal segments) that match the behavioral signal to be detected within the selected category of behavioral signals. For example, red lines can be used to highlight these two signal segments, and corresponding text descriptions and explanations can be added to the user interface, such as "During the abnormally high heart rate, the patient's blood pressure also increased significantly, while the number of steps decreased significantly."

[0294] Through the above steps, the computer system can accurately retrieve multiple signal segments that match the detected behavioral signal from the selected category of behavioral signals, and determine which signal segments should be emphasized by evaluating the continuity between these signal segments.

[0295] In one implementation, step S1421, based on the signal coordinates of the third signal segment and the fourth signal segment in the selected category of behavioral signals, determines the continuity evaluation parameters between the third signal segment and the fourth signal segment, including:

[0296] Step S14211: Determine the signal spacing between the third signal segment and the fourth signal segment based on the signal coordinates of the third signal segment in the selected category of behavioral signals and the signal coordinates of the fourth signal segment in the selected category of behavioral signals;

[0297] Step S14212: Determine the continuity evaluation parameters between the third and fourth signal segments based on the signal coordinate reference of the third and fourth signal segments and the spacing between the third and fourth signal segments.

[0298] In step S14211, the computer system calculates the signal spacing between the third and fourth signal segments. Signal spacing is a quantitative indicator used to measure the distance between two signal segments in time or space. For time-series data, signal spacing typically represents the interval between two signal segments on the time axis; for non-time-series data, signal spacing may represent the distance between two signal segments in the feature space.

[0299] If the third and fourth signal segments are two time periods in a time series data set, the computer system can calculate the difference between the start and end times of these two time periods to obtain the signal interval. For example, if the start time of the third signal segment is t1 and the end time is t2, and the start time of the fourth signal segment is t3 and the end time is t4, then the signal interval can be calculated as |t3 + t2| (assuming the two signal segments are continuous in time and do not overlap).

[0300] For non-time-series data, such as images and text, computer systems can first convert signal segments into feature vectors, and then use metrics such as Euclidean distance, Manhattan distance, or cosine similarity to calculate the signal spacing. For example, if the third and fourth signal segments are converted into feature vectors v1 and v2 respectively, then the signal spacing can be calculated as ‖v1 v2‖ (Euclidean distance).

[0301] For example, suppose we are analyzing the heart rate and blood glucose data of a diabetic patient. The third signal segment is a time period within the heart rate data, representing rapid changes in the patient's heart rate over a given time period; the fourth signal segment is a time period within the blood glucose data, representing corresponding changes in the patient's blood glucose level. If both signal segments are time-series data, the computer system can calculate the time interval between them as the signal gap. For example, if the third signal segment starts at 9:00 AM and ends at 9:10 AM, and the fourth signal segment starts at 9:05 AM and ends at 9:15 AM, then the signal gap is 5 minutes.

[0302] After determining the signal interval, the computer system considers the signal coordinate reference value of the third and fourth signal segments, as well as the signal interval itself, to comprehensively evaluate the continuity between these two signal segments. The computer system can assign a reference score to each signal segment, reflecting its importance in the analysis. The reference score can be determined based on the characteristics of the signal segments (such as mean, variance, kurtosis, etc.), patterns (such as periodicity, trends, etc.), or outliers (such as extreme values, abrupt changes, etc.). For example, if the third signal segment contains a significant heart rate peak, and the fourth signal segment also shows a corresponding blood glucose peak, then the reference scores for both signal segments can be high.

[0303] When calculating continuity evaluation parameters, the computer system can weight the signal spacing to take into account the reference value of the signal coordinates. Specifically, if both signal segments have high reference values, the continuity evaluation parameter may be high even if the signal spacing between them is slightly large; conversely, if the signal spacing is small but the reference value is low, the continuity evaluation parameter can be low.

[0304] The continuity assessment parameters can be calculated using the following formula:

[0305] ;

[0306] The reciprocal of the signal spacing is used to convert the spacing into a quantity positively correlated with continuity (because a smaller spacing indicates better continuity), while the sum of the reference scores reflects the importance of the two signal segments in the analysis. This formula can be adjusted according to actual needs, for example, by introducing a weighting factor to balance the influence of the signal spacing and the reference score.

[0307] For example, let's continue using heart rate and blood glucose data from diabetic patients. Assume the reference scores for the third and fourth signal segments are 8 and 9 (out of 10), respectively, with a signal interval of 5 minutes. Based on the formula above, the continuity assessment parameter can be calculated as follows:

[0308] ;

[0309] This value indicates good continuity between the third and fourth signal segments. Although the time interval between them is 5 minutes, the continuity evaluation parameter is still high because their reference scores are high (indicating that these two signal segments are very important in the analysis).

[0310] Through the above steps, the computer system can accurately calculate the signal spacing between the third and fourth signal segments, and determine the continuity evaluation parameters by comprehensively considering the reference value of the signal coordinates and the signal spacing. This not only helps to evaluate the tightness and continuity between the two signal segments, but also provides strong support for subsequent data analysis and pattern recognition.

[0311] This application provides a computer system, such as... Figure 2 As shown, the computer system 100 includes a processor 101 and a memory 103. The processor 101 and the memory 103 are connected, for example, via a bus 102. Optionally, the computer system 100 may also include a transceiver 104. It should be noted that in practical applications, the transceiver 104 is not limited to one type, and the structure of this computer system 100 does not constitute a limitation on the embodiments of this application.

[0312] This application provides a computer system, which includes: one or more processors; a memory; and one or more computer programs, wherein the one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more programs are executed by the processors, they implement the above-mentioned method for analyzing patient signal data from monitoring sources.

[0313] The above description is only a partial embodiment of this application. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the principle of this application, and these improvements and modifications should also be considered within the scope of protection of this application.

Claims

1. A method for analyzing patient signal data from a monitoring source, characterized in that, The method includes: Acquire target patient behavior signals that are compatible with the behavior signals to be detected, wherein the target patient behavior signals cover different categories of behavior signals; Based on the influence factors of the different categories of behavioral signals, the matching degree between the different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected is obtained; Based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected, a selected category of behavioral signals is obtained by filtering from the different categories of behavioral signals; The signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are highlighted and displayed. The step of highlighting and displaying the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected includes: Multiple signal segments to be emphasized are retrieved from the selected category of behavioral signals and are adapted to the behavioral signal to be detected; The continuity evaluation parameters among the multiple signal segments to be emphasized are determined by the signal coordinates of the multiple signal segments to be emphasized in the behavioral signal of the selected category; When the continuity evaluation parameters among the multiple signal segments to be emphasized reach the preset standard, the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected are emphasized and displayed.

2. The method as described in claim 1, characterized in that, The step of obtaining the matching degree between the different categories of behavioral signals covered by the target patient's behavioral signals and the behavioral signals to be detected, based on the influence factors of the different categories of behavioral signals, includes: The signal to be detected is segmented into signal segments to obtain the signal segments to be detected; Based on the signal segment to be detected, the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected is obtained; Based on the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected, and the influence factor of the different categories of behavioral signals, the matching degree between the different categories of behavioral signals and the behavioral signal to be detected is obtained.

3. The method as described in claim 2, characterized in that, The step of obtaining the basic matching degree between the different categories of behavioral signals and the behavioral signals to be detected based on the signal segment to be detected includes: Target signal segments that match the signal segment to be detected are retrieved from the different categories of behavioral signals; Based on the priority of the signal segment to be detected and the priority of the target signal segment that matches the signal segment to be detected among the different categories of behavioral signals, the basic matching degree between the different categories of behavioral signals and the behavioral signal to be detected is determined.

4. The method as described in claim 1, characterized in that, The method includes multiple target patient behavioral signals and further includes: Based on the adaptation mechanism between the behavioral signal to be detected and the behavioral signal of each target patient, the influence factor of each target patient's behavioral signal is determined; Based on the influence factor of each target patient behavior signal and the matching degree between the different categories of behavior signals covered by each target patient behavior signal and the behavior signal to be detected, the matching degree between each target patient behavior signal and the behavior signal to be detected is obtained; Based on the matching degree between each target patient behavior signal and the behavior signal to be detected, multiple target patient behavior signals are listed in sequence. The acquisition of the target patient behavioral signal that matches the behavioral signal to be detected includes: The signal to be detected is segmented into signal segments to obtain a set of signal segments to be detected; The signal segments in the set of signal segments to be detected are divided into clusters to obtain multiple signal segment clusters; Target patient behavior signals that are adapted to the multiple signal segment clusters are acquired respectively, with each signal segment cluster corresponding to one adaptation mechanism.

5. The method as described in claim 4, characterized in that, The step of segmenting the target behavior signal into signal segments to obtain a set of target signal segments includes: The behavior patterns of the detected behavior signals are classified to obtain behavior pattern labels for the detected behavior signals; The signal to be detected is segmented, and the set of signal segments to be detected is generated based on the behavior pattern label and the signal segments obtained from the signal segmentation.

6. The method as described in claim 1, characterized in that, The step of filtering out selected category behavior signals from the different categories of behavior signals based on the matching degree between the different categories of behavior signals and the behavior signal to be detected includes: The target category of behavioral signals is obtained by filtering from the different categories of behavioral signals, as well as behavioral signals of unselected categories other than the target category; Based on the matching degree between the different categories of behavioral signals and the behavioral signals to be detected, behavioral signals of undetermined categories are filtered from the behavioral signals of unselected categories. The behavioral signals of the undetermined category and the behavioral signals of the target category are used as the behavioral signals of the selected category; The method of highlighting and displaying signal segments in the selected category of behavioral signals that match the behavioral signal to be detected also includes: The target category signal segments are obtained by filtering from the set of signal segments to be detected, and the set of signal segments to be detected includes signal segments obtained by segmenting the signal segments of the behavior signal to be detected. The signal segments in the selected category of behavioral signals that match the signal segments of the target category are highlighted and displayed.

7. The method as described in claim 6, characterized in that, The step of filtering behavioral signals of unselected categories from the behavioral signals based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected includes: Based on the matching degree between the different categories of behavioral signals and the behavioral signal to be detected, the original undetermined category of behavioral signals with a matching degree higher than the matching degree threshold is selected from the behavioral signals of unselected categories. The behavioral signals of the original undetermined categories corresponding to the top K matching degrees in descending order are taken as the behavioral signals of the undetermined categories, where K≥1.

8. The method as described in claim 7, characterized in that, The step of highlighting and displaying the signal segments in the selected category of behavioral signals that match the behavioral signal to be detected also includes: Identify unselected category signal segments in the set of signal segments to be detected, other than those of the target category. A candidate signal segment set is obtained by retrieving signal segments that match the unselected category signal segments from the selected category of behavioral signals. G signal segments are selected from the candidate signal segment set and highlighted for display, where G≥1.

9. A computer system, characterized in that, include: One or more processors; Memory; One or more computer programs; The one or more computer programs are stored in the memory and configured to be executed by the one or more processors, and when the one or more computer programs are executed by the processors, they implement the method as described in any one of claims 1 to 8.

Images