CN114334090B - Data analysis method and device and electronic equipment - Google Patents
Data analysis method and device and electronic equipment Download PDFInfo
- Publication number
- CN114334090B CN114334090B CN202210196265.8A CN202210196265A CN114334090B CN 114334090 B CN114334090 B CN 114334090B CN 202210196265 A CN202210196265 A CN 202210196265A CN 114334090 B CN114334090 B CN 114334090B
- Authority
- CN
- China
- Prior art keywords
- data
- data analysis
- preset
- initial
- heart rhythm
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Abstract
The invention provides a data analysis method, a data analysis device and electronic equipment. Because the preset data analysis model is obtained based on a large amount of training data, the preset data analysis model can accurately identify the data analysis result of the representation emotion state of the target user, and the analysis accuracy and efficiency requirements on the data representing emotion are met. Furthermore, the method can be applied to a scene of recognizing the emotional state of the user, and the data analysis result of the method is used for assisting the recognition of the emotional state of the user so as to meet the requirement of recognizing the emotional state of the user in daily life.
Description
Technical Field
The invention relates to the field of data processing and analysis, in particular to a data analysis method and device and electronic equipment.
Background
The emotion is a state integrating human feelings, ideas and behaviors, and plays an important role in human-to-human communication. The mood plays a ubiquitous role in the daily work and life of people. If the emotional state of the patient can be known, as in medical care, different care measures can be taken according to the emotion of the patient to improve the care quality.
In particular, anxiety and depression emotions are a psychological health problem, and people in all age groups with anxiety and depression are in a rapid growth trend in recent years, so that society pays more and more attention to the psychological health of people. Particularly, a large number of people need to be isolated at home due to epidemic situations, anxiety of different degrees or generalized anxiety may occur to the public during the isolation period, and if the anxiety mood cannot be timely relieved, serious adverse consequences such as depression and the like easily develop. Therefore, psychological consultants and psychiatrists are invested in various health institutions all over the country during epidemic outbreaks to analyze the characteristic emotion data, or psychological health dispersion services are set up on the internet or radio stations, so that the risk of cross infection caused by hospitals is reduced as much as possible.
At present, when data representing emotion is analyzed, an analysis result is mainly obtained in a mode of manually analyzing a scale filled by a user. The data analysis mode has low efficiency and accuracy and cannot meet the data analysis requirement.
Disclosure of Invention
In view of this, the present invention provides a data analysis method, an apparatus and an electronic device, so as to solve the problem that the efficiency and accuracy of analyzing data representing emotion are low and the data analysis requirement cannot be met.
In order to solve the technical problems, the invention adopts the following technical scheme:
a method of data analysis, comprising:
acquiring initial heart rhythm data of a target user, and performing preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data;
calling a preset data analysis model to perform data processing on the target heart rhythm data so as to obtain a data analysis result corresponding to the target heart rhythm data; the preset data analysis model is obtained based on training data; the training data comprises a heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample; the data analysis result is used for representing emotional states;
and under the condition that the data analysis result is a preset data analysis result, acquiring data analysis associated information, and under the condition that a data query request sent by the terminal of the target user is received, outputting the data analysis result and the data analysis associated information to the terminal.
Optionally, obtaining initial heart rhythm data of the target user comprises:
acquiring an initial heart rhythm data set stored in a preset storage position; the initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data;
and acquiring target user information of a target user, and screening initial rhythm data corresponding to the target user information from the initial rhythm data set according to the user information corresponding to the initial rhythm data.
Optionally, performing preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data, including:
performing data cleaning, data sorting and data downsampling operation on the initial heart rhythm data to obtain heart rhythm data to be processed;
and determining target rhythm data corresponding to the rhythm data to be processed according to a preset data processing mode.
Optionally, determining target heart rate data corresponding to the to-be-processed heart rate data according to a preset data processing manner includes:
taking the heart rhythm data to be processed as target heart rhythm data;
or extracting time domain characteristic information and frequency domain characteristic information of the heart rhythm data to be processed;
and sequentially combining the heart rhythm data to be processed, the time domain characteristic information and the frequency domain characteristic information to obtain target heart rhythm data.
Optionally, the generating process of the preset data analysis model includes:
acquiring initial data samples arranged in a long-time sequence, and performing preset data processing operation on the initial data samples to obtain rhythm data samples;
acquiring a data analysis result label of the initial data sample, and using the data analysis result label as a data analysis result label of the heart rhythm data sample;
and training a preset data analysis model by using the heart rhythm data sample and the data analysis result label corresponding to the heart rhythm data sample until a preset training stop condition is met.
Optionally, after a preset data analysis model is called to perform data processing on the target heart rhythm data to obtain a data analysis result corresponding to the target heart rhythm data, the method further includes:
acquiring data analysis results of the continuous preset number of the target users;
determining the number of data analysis results as preset data analysis results, and calculating the ratio of the number to the total number of the data analysis results;
and taking a data analysis result corresponding to the preset ratio interval in which the ratio is positioned as a data analysis correction result of the target user.
A data analysis apparatus comprising:
the data processing module is used for acquiring initial heart rhythm data of a target user and carrying out preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data;
the data analysis module is used for calling a preset data analysis model to perform data processing on the target heart rhythm data so as to obtain a data analysis result corresponding to the target heart rhythm data; the preset data analysis model is obtained based on training data; the training data comprises a heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample; the data analysis result is used for representing emotional states;
and the data output module is used for acquiring data analysis associated information under the condition that the data analysis result is a preset data analysis result, and outputting the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received.
Optionally, the data processing module includes:
the set acquisition submodule is used for acquiring an initial heart rhythm data set stored in a preset storage position; the initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data;
and the data screening submodule is used for acquiring target user information of a target user and screening initial heart rhythm data corresponding to the target user information from the initial heart rhythm data set according to the user information corresponding to the initial heart rhythm data.
Optionally, the data processing module includes:
and the data processing submodule is used for carrying out data cleaning, data sorting and data down-sampling operations on the initial heart rhythm data to obtain the heart rhythm data to be processed, and determining target heart rhythm data corresponding to the heart rhythm data to be processed according to a preset data processing mode.
An electronic device, comprising: a memory and a processor;
wherein the memory is used for storing programs;
the processor calls a program and is used to perform the data analysis method described above.
Compared with the prior art, the invention has the following beneficial effects:
the invention provides a data analysis method, a data analysis device and electronic equipment. The preset data analysis model is obtained based on a large amount of training data through training, the training data comprise rhythm data samples and data analysis result labels corresponding to the rhythm data samples, and therefore the preset data analysis model can accurately identify the data analysis results representing the emotion states of the target user, and the requirements for the analysis accuracy and efficiency of the data representing emotion are met. In addition, the data analysis method and the data analysis system can also acquire data analysis related information under the condition that the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis related information to the terminal under the condition that a data query request sent by the terminal of the target user is received. Namely, the invention supports data query, and sends the corresponding data analysis associated information to the user during query, so as to improve the user experience. Furthermore, the data analysis method, the data analysis device and the electronic equipment can assist in recognizing the emotional state according to the data analysis result of the invention.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the provided drawings without creative efforts.
FIG. 1 is a flowchart of a method for analyzing data according to an embodiment of the present invention;
FIG. 2 is a flow chart of another method for data analysis according to an embodiment of the present invention;
FIG. 3 is a flowchart of a method of analyzing data according to another embodiment of the present invention;
FIG. 4 is a flowchart of a method for analyzing data according to another embodiment of the present invention;
fig. 5 is a schematic structural diagram of a data analysis apparatus according to an embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The emotion is a state integrating human feelings, ideas and behaviors, and plays an important role in human-to-human communication. The mood plays a ubiquitous role in the daily work and life of people. If the emotional state of the patient can be known, as in medical care, different care measures can be taken according to the emotion of the patient to improve the care quality.
At present, when data representing emotion is analyzed, an analysis result is mainly obtained in a mode of manually analyzing a scale filled by a user. The data analysis mode has low efficiency and accuracy and cannot meet the data analysis requirement.
In analyzing the data representing emotions, such as analyzing the depressed mood data, manually analyzing the scales filled by the user, such as the zoned depression self-rating scale (SDS), the hamilton depression scale (HRSD), the depression scale (PHQ-9), the hamilton anxiety scale (HAMA), the generalized anxiety scale (GAD-7), and the zoned anxiety self-rating scale (SAS), the doctor needs to classify according to different states of the subject and then fill different scales for different subjects, and the whole process of filling the scales takes a long time and needs guidance of the doctor. Furthermore, the answers to the scale are subject to subject compliance and authenticity, and the subjective bias of the results is strong, thereby making the analysis of depressed mood data less accurate.
In order to improve the analysis accuracy of anxiety and depression emotion recognition data, images and audios of a subject can be used for recognition, but the defects that privacy protection of the subject is involved, image and audio data are difficult to collect and have low popularity, doctors are difficult to label data and the like are achieved, and the progress is slow.
HRV is the abbreviation of heart rate variability, HRV is a good index for observing and knowing body changes, Heart Rate Variation (HRV) represents such a quantitative mapping, namely HRV reflects the degree and rule of heart rate change by measuring the HRV of continuous normal R-R interval change, the heart rate variation means that the human heart beating cycle has tiny variation, in recent years, HRV is widely applied as a means for non-invasively detecting the heart function, HRV signals contain a large amount of information concerning the regulation of nerves and body fluids of a vascular system, and a plurality of information in the aspect of autonomic nervous system can be obtained by analyzing heart rate data, and the information can be extracted to quantitatively evaluate the tensity and balance of sympathetic nerves of the heart and the influence of the sympathetic nerves on the activity of the cardiovascular system.
At present, HRV is mainly used for researching diseases closely related to autonomic nerves, such as diabetic neuropathy, cardiovascular and cerebrovascular diseases, arteriosclerosis, neurasthenia, heart failure, insomnia, depression, anxiety and the like in medicine, and the research mainly calculates time domain, frequency domain and nonlinear dynamics characteristic indexes, wherein the time domain comprises: SDNN (24 hRR interval standard deviation), SDANN (standard deviation of 24h per 5minRR interval mean), RMSSD (root mean square of 24h adjacent RR interval difference), PNN50 (NN 50 divided by total NN interval number), frequency domain comprising: indexes such as ULF (ultra high frequency power), VLF (ultra low frequency power), LF (low frequency power), HF (high frequency power), LF/HF (balance ratio) and the like are utilized to have some exploratory researches on depression and generalized anxiety disorder, and the analysis of anxiety and depression mood data is mainly carried out by manually analyzing time domain, frequency domain, nonlinear dynamics characteristic indexes and the like of HRV.
However, this method is easily affected by artificial subjective factors, and the analysis accuracy of anxiety and depression data is low.
In the invention, initial rhythm data of a target user are obtained, preset data processing operation is carried out on the initial rhythm data to obtain target rhythm data, a preset data analysis model is called to carry out data processing on the target rhythm data to obtain a data analysis result corresponding to the target rhythm data, and the data analysis result is used for representing an emotional state. The preset data analysis model is obtained based on a large amount of training data through training, the training data comprise rhythm data samples and data analysis result labels corresponding to the rhythm data samples, and therefore the preset data analysis model can accurately identify the data analysis results representing the emotion states of the target user, and the requirements for the analysis accuracy and efficiency of the data representing emotion are met. In addition, the data analysis method and the data analysis system can also acquire data analysis associated information under the condition that the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received. Namely, the invention supports data query, and sends the corresponding data analysis associated information to the user during query, so as to improve the user experience.
On the basis of the above, another embodiment of the present invention provides a data analysis method, and with reference to fig. 1, the data analysis method may include:
and S11, acquiring initial rhythm data of the target user, and performing preset data processing operation on the initial rhythm data to obtain target rhythm data.
In this embodiment, the initial heart rate data may be HRV data, and may be acquired by a user wearing a watch.
The initial cardiac rhythm data in this embodiment is a long-term time sequence signal, specifically, a continuous time sequence RR interval heartbeat signal exceeding 24 hours. In this embodiment, the long-time sequence signal is used for prediction, which is superior to the short-time sequence signal prediction in that short-time monitoring can only monitor data in a short time, which time interval is abnormal cannot be determined, and the long-time data can capture the abnormal wave band in 24 hours.
In practical applications, acquiring initial heart rate data of a target user may include:
1) and acquiring an initial heart rhythm data set stored in a preset storage position.
The initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data.
Specifically, the user registers user information (such as a user ID number, a mobile phone number and the like) at the watch end, then the user wears the watch according to the requirement, the watch can automatically upload initial heart rate data to the watch cloud database in real time, and the user logs in the user name registered at the watch end by using mobile phone software APP, so that current and historical HRV data, specifically pulse wave RR interval data, can be observed.
Because there are more users to upload initial rhythm of heart data, so, after wrist-watch end user registers the ID, the data that the wrist-watch was gathered can be automatically uploaded wrist-watch high in the clouds database, and initial rhythm of heart data set is stored to the high in the clouds database, initial rhythm of heart data set includes the initial rhythm of heart data that intelligent wearing equipment that a plurality of users wore gathered and uploaded and user information that initial rhythm of heart data corresponds. The intelligent wearable device is the watch.
2) And acquiring target user information of a target user, and screening initial rhythm data corresponding to the target user information from the initial rhythm data set according to the user information corresponding to the initial rhythm data.
In this embodiment, target user information of a target user, such as a user ID number and a mobile phone number, is obtained, and then initial heart rate data corresponding to the target user information is screened out from the initial heart rate data set according to the user information corresponding to the initial heart rate data.
After initial heart rate data of a target user are acquired, preset data processing operation is carried out on the initial heart rate data to obtain target heart rate data.
Specifically, referring to fig. 2, performing preset data processing operation on the initial heart rate data to obtain target heart rate data includes:
and S21, performing data cleaning, data sorting and data down-sampling operation on the initial heart rhythm data to obtain the heart rhythm data to be processed.
And carrying out data desensitization on the pulse wave RR interval data, namely carrying out desensitization operation on some important data such as mobile phone numbers and the like for data cleaning operation. Then, the data is rearranged according to the sequence from 0:00 hour to 24:00 hour, and the regularity of the data is ensured.
After which a data down-sampling operation is performed. The sampling frequency of the watch is 100HZ, 70000-100000 heartbeat counts in 24 hours, the sampling step length is 10 heartbeat counts, and 7000-10000 data points are obtained from normal 24-hour data. 7000-10000 data points are the data of the heart rhythm to be processed.
And S22, determining target rhythm data corresponding to the rhythm data to be processed according to a preset data processing mode.
Specifically, the heart rate data to be processed may be directly used as the target heart rate data. In addition, in order to improve the reliability of the data, the time domain characteristic information and the frequency domain characteristic information of the heart rhythm data to be processed can be extracted. Wherein the time domain feature information includes:
SDNN (24 hRR interval standard deviation), SDANN (standard deviation of mean of 24h intervals per 5 minRR), RMSSD (root mean square of difference between adjacent RR intervals of 24 h), PNN50 (NN 50 divided by total NN interval number).
The frequency domain characteristic information includes:
indexes such as ULF (ultra high frequency power), VLF (ultra low frequency power), LF (low frequency power), HF (high frequency power), and LF/HF (balance ratio).
And then sequentially combining the heart rhythm data to be processed, the time domain characteristic information and the frequency domain characteristic information to obtain target heart rhythm data.
The data obtained in the mode is subjected to model training, so that the model can learn the whole data and can also learn the time domain characteristic information and the frequency domain characteristic information in a key mode.
And S12, calling a preset data analysis model to perform data processing on the target heart rhythm data so as to obtain a data analysis result corresponding to the target heart rhythm data.
In this embodiment, the preset data analysis model is obtained by training based on training data; the training data comprises a heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample; and the data analysis result is used for representing the emotional state. The emotional state may be an anxious depressed mood or a non-anxious depressed mood.
In practical applications, a preset data analysis model is generated in advance, and referring to fig. 3, a generation process of the preset data analysis model includes:
and S31, obtaining initial data samples arranged in a long-term time sequence, and carrying out preset data processing operation on the initial data samples to obtain the heart rhythm data samples.
The model generation process in this embodiment includes two stages, the first stage is a model data acquisition stage, data mainly comes from the department of outpatient service of the institute of traditional Chinese medicine science, the data acquisition time is mainly concentrated in the epidemic situation period, in order to reduce the risk of cross infection in the hospital, 325 pieces of data of 91 cases of depression and anxiety patients are acquired by adopting a mode of sending a watch to the patients, the data acquisition distribution is shown in table 1, wherein 97% of data acquisition is concentrated in the period from 7 months to 1 month of 2021 in 2020. 400 pieces of data of 191 normal persons are collected at the same time. In this embodiment, the data collected is an initial data sample with a long time sequence arrangement.
TABLE 1
| Time of acquisition | 2018.01 | 2020.07 | 2020.08 | 2020.09 | 2020.10 | 2020.12 | 2021.01 | 2021.04 | 2021.05 |
| Number of collected | 1 | 3 | 74 | 26 | 2 | 92 | 119 | 8 | 1 |
After the data are acquired from the watch end cloud, the data are used as initial data samples, and then preset data processing operation is carried out on the initial data samples to obtain the heart rhythm data samples.
The specific process of the preset data processing operation refers to the corresponding process.
And S32, acquiring a data analysis result label of the initial data sample, and using the data analysis result label as the data analysis result label of the heart rhythm data sample.
Step S32 is the first stage of model generation. After the above-mentioned heart rhythm data samples are acquired, the data analysis result labels of the heart rhythm data samples also need to be determined. If the initial data sample corresponding to the heart rhythm data sample is the data of the patient with the melancholia and anxiety, the data analysis result label of the initial data sample is the anxious and depressed mood, and at the moment, the data analysis result label of the heart rhythm data sample is also set as the anxious and depressed mood.
And if the initial data sample corresponding to the heart rhythm data sample is the data of the normal person, the data analysis result label of the initial data sample is the non-anxious depressed mood, and at the moment, the data analysis result label of the heart rhythm data sample is also set to be the non-anxious depressed mood.
And S33, training a preset data analysis model by using the rhythm data sample and the data analysis result label corresponding to the rhythm data sample until a preset training stop condition is met.
Step S33 is the second stage of model generation. The preset data analysis model in this embodiment may be a binary classification model trained by a time-series random forest classification algorithm, the principle of the random forest classification algorithm is to classify according to a plurality of tickets of all tree branches in a forest, and in the decision of the plurality of tickets, the prediction result is determined by the classification result predicted by the most tree branches. The time series forest is an interpretable model, and the importance of time features can be extracted from the time series forest.
The time sequence random forest classifier firstly divides sequence data into random intervals, random initial positions and random lengths, extracts features from each section, trains a decision tree by using the features, and continuously splits tree nodes until leaves contain minimum sample numbers, and each tree can find the best feature. Specifically, the training process is as follows:
1) dividing the sequence into random intervals, random initial positions and random lengths;
2) extracting features (mean, standard deviation, slope) from each interval into a single feature vector;
3) a decision tree is trained on the extracted features.
And (4) repeating the steps 1-3 until the number of the trees required to be established or the time is exhausted, and stopping training when the preset training stopping condition is met.
During training, the ratio of 7: 3, training and testing of the long-term time sequence random forest classification model are carried out in proportion, the training model adopts a ten-fold cross validation method, and through test tests, the analysis accuracy of depression anxiety data reaches 90%, the analysis accuracy of non-anxiety depression data reaches 87%, and the accuracy of the questionnaire scheme is far exceeded.
And after the preset data analysis model is generated, deploying the preset data analysis model to a cloud server. The watch data are uploaded regularly when a user wears the watch, after 24-hour data are collected accumulatively, the cloud server obtains data from a watch cloud database according to user information such as a user ID number and a mobile phone number, then a preset data analysis model is called to process the collected data, a data analysis result is obtained, the data analysis result is used for representing the emotional state, and the specific content of the data analysis result is anxiety and depression emotion or non-anxiety and depression emotion. In addition, anxious and depressive scores may also be given.
The preset data analysis model in this embodiment may also adopt other machine learning methods, such as logistic regression, K nearest neighbor of time series, and the like.
And S13, acquiring data analysis associated information under the condition that the data analysis result is a preset data analysis result, and outputting the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received.
In this embodiment, the preset data analysis result may represent data with anxiety and depression mood. And in the case that the data analysis result is a preset data analysis result, indicating that the user has an anxious and depressed mood. At this time, data analysis related information is acquired, and the data analysis related information may be a series of health conditioning protocols for depression and anxiety, specifically referring to table 2.
TABLE 2
The target user can log in the mobile phone APP at any time to inquire the emotion analysis result of the target user. The mobile phone APP needs to register the mobile phone number of the user, the mobile phone number is bound with the watch to be the same, and the testee can watch the prediction result after taking the watch for more than 24 hours.
At this time, the cloud server receives a data query request sent by the terminal of the target user, and outputs the data analysis result and the data analysis associated information to the terminal. That is, in this embodiment, when the user has an anxious and depressed mood, the data analysis result is transmitted to the user, and the related health conditioning scheme is also transmitted to the user, so that the user can know in time how to eliminate the anxious and depressed mood.
In practical application, the emotion of the user may fluctuate greatly due to emergencies such as illness, and the emotion recognition result of certain data cannot represent the long-time emotion state of the user. For example, a user's mood is always stable, but the mood fluctuates greatly due to illness, and spleen qi is likely to occur.
In order to determine the long-time emotional state of the user, in this embodiment, referring to fig. 4, after invoking a preset data analysis model to perform data processing on the target heart rate data to obtain a data analysis result corresponding to the target heart rate data, the method further includes:
and S41, acquiring the data analysis results of the continuous preset number of the target users.
The continuous preset number in this embodiment may be one week, that is, 7 continuous data analysis results.
S42, determining the number of the data analysis results as preset data analysis results, and calculating the ratio of the number to the total number of the data analysis results.
That is, in this embodiment, the preset data analysis results are screened out from all the data analysis results, the number of the preset data analysis results is counted, and then the ratio of the number to the total number of the data analysis results is calculated.
And S43, taking the data analysis result corresponding to the preset ratio interval in which the ratio is positioned as the data analysis and correction result of the target user.
Specifically, a data analysis result corresponding to the preset ratio interval may be preset, for example:
the preset ratio interval is 0.8-1, and the data analysis result shows that the anxiety and depression mood is obtained;
the preset ratio interval is 0.5-0.8, and the data analysis result shows moderate anxiety and depression emotion;
the preset ratio interval is 0.3-0.5, and the data analysis result shows mild anxiety and depression;
the preset ratio interval is 0-0.3, and the data analysis result is non-anxiety depression mood.
And after the ratio is determined, determining a preset ratio interval where the ratio is located, and then taking a data analysis result corresponding to the preset ratio interval as a data analysis correction result of the target user. By this step, the long-term emotional state of the user can be determined.
By utilizing the data analysis result in the scheme, the emotional state can be identified in an auxiliary manner, and the emotional state of the testee can be predicted in an auxiliary manner under the state of daily life.
In this embodiment, initial heart rhythm data of a target user is obtained, preset data processing operation is performed on the initial heart rhythm data to obtain target heart rhythm data, a preset data analysis model is called to perform data processing on the target heart rhythm data to obtain a data analysis result corresponding to the target heart rhythm data, and the data analysis result is used for representing an emotional state. The preset data analysis model is obtained based on a large amount of training data through training, the training data comprise rhythm data samples and data analysis result labels corresponding to the rhythm data samples, and therefore the preset data analysis model can accurately identify the data analysis results representing the emotion states of the target user, and the requirements for the analysis accuracy and efficiency of the data representing emotion are met. In addition, the data analysis method and the data analysis system can also acquire data analysis related information under the condition that the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis related information to the terminal under the condition that a data query request sent by the terminal of the target user is received. Namely, the invention supports data query, and sends the corresponding data analysis associated information to the user during query, so as to improve the user experience.
In addition, the user only needs to carry the monitoring watch, normal life of 24 hours is not affected, and the watch system automatically uploads heart rate data to the cloud; the cloud server performs preprocessing and analysis according to the heart rate data of the long-term time sequence of the user, so that the depression and anxiety prediction degree is given, and the privacy of the user is well protected.
In addition, the invention utilizes the long-term time sequence signal data to carry out classification prediction based on a machine learning model, replaces the test subject to fill in a questionnaire scale or audio-video monitoring, and has more objective data.
In addition, the invention gives a personalized health conditioning scheme according to the predicted result of the data of the subjects.
In addition, the invention can assist self-test whether the anxiety and depression tend to exist, and if the anxiety and depression tend to exist, the self-regulation can be carried out according to the health management scheme provided by the system, the emotional stress and the social stress can be relieved.
Alternatively, on the basis of the above-mentioned embodiment of the data analysis method, another embodiment of the present invention provides a data analysis apparatus, referring to fig. 5, including:
the data processing module 11 is configured to obtain initial heart rate data of a target user, and perform preset data processing operation on the initial heart rate data to obtain target heart rate data;
the data analysis module 12 is configured to call a preset data analysis model to perform data processing on the target heart rate data to obtain a data analysis result corresponding to the target heart rate data; the preset data analysis model is obtained based on training data; the training data comprises a heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample; the data analysis result is used for representing the emotional state;
the data output module 13 is configured to acquire data analysis associated information when the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis associated information to the terminal when a data query request sent by the terminal of the target user is received.
Further, the data processing module comprises:
the set acquisition submodule is used for acquiring an initial heart rhythm data set stored in a preset storage position; the initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data;
and the data screening submodule is used for acquiring target user information of a target user and screening initial heart rhythm data corresponding to the target user information from the initial heart rhythm data set according to the user information corresponding to the initial heart rhythm data.
Further, the data processing module includes:
and the data processing submodule is used for carrying out data cleaning, data sorting and data down-sampling operations on the initial heart rhythm data to obtain the heart rhythm data to be processed, and determining target heart rhythm data corresponding to the heart rhythm data to be processed according to a preset data processing mode.
Further, the data processing sub-module is configured to, when determining target rhythm data corresponding to the to-be-processed rhythm data according to a preset data processing manner, specifically:
taking the heart rhythm data to be processed as target heart rhythm data;
or extracting time domain characteristic information and frequency domain characteristic information of the heart rhythm data to be processed, and sequentially combining the heart rhythm data to be processed, the time domain characteristic information and the frequency domain characteristic information to obtain target heart rhythm data.
Further, the method also comprises a model generation module, which comprises the following steps:
the sample acquisition submodule is used for acquiring initial data samples arranged in a long-term time sequence and carrying out preset data processing operation on the initial data samples to obtain rhythm data samples;
the label obtaining submodule is used for obtaining a data analysis result label of the initial data sample and using the data analysis result label as the data analysis result label of the heart rhythm data sample;
and the training submodule is used for training a preset data analysis model by using the heart rhythm data sample and the data analysis result label corresponding to the heart rhythm data sample until a preset training stop condition is met.
Further, still include:
the result acquisition module is used for acquiring data analysis results of the continuous preset number of the target users;
the ratio calculation module is used for determining the number of data analysis results as preset data analysis results and calculating the ratio of the number to the total number of the data analysis results;
and the result correction module is used for taking the data analysis result corresponding to the preset ratio interval in which the ratio is positioned as the data analysis correction result of the target user.
In this embodiment, initial heart rhythm data of a target user is obtained, preset data processing operation is performed on the initial heart rhythm data to obtain target heart rhythm data, a preset data analysis model is called to perform data processing on the target heart rhythm data to obtain a data analysis result corresponding to the target heart rhythm data, and the data analysis result is used for representing an emotional state. The preset data analysis model is obtained based on a large amount of training data through training, the training data comprise rhythm data samples and data analysis result labels corresponding to the rhythm data samples, and therefore the preset data analysis model can accurately identify the data analysis results representing the emotion states of the target user, and the requirements for the analysis accuracy and efficiency of the data representing emotion are met. In addition, the data analysis method and the data analysis system can also acquire data analysis associated information under the condition that the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received. The invention supports data query, and sends the corresponding data analysis associated information to the user during query, so as to improve the user experience.
It should be noted that, for the working processes of each module and sub-module in this embodiment, please refer to the corresponding description in the above embodiments, which is not described herein again.
Optionally, on the basis of the embodiments of the data analysis method and apparatus, another embodiment of the present invention provides an electronic device, including: a memory and a processor;
wherein the memory is used for storing programs;
the processor calls a program and is used to perform the data analysis method described above.
In the embodiment, initial heart rate data of a target user is obtained, preset data processing operation is performed on the initial heart rate data to obtain target heart rate data, a preset data analysis model is called to perform data processing on the target heart rate data to obtain a data analysis result corresponding to the target heart rate data, and the data analysis result is used for representing an emotional state. The preset data analysis model is obtained based on a large amount of training data through training, the training data comprise rhythm data samples and data analysis result labels corresponding to the rhythm data samples, and therefore the preset data analysis model can accurately identify the data analysis results representing the emotion states of the target user, and the requirements for the analysis accuracy and efficiency of the data representing emotion are met. In addition, the data analysis method and the data analysis system can also acquire data analysis associated information under the condition that the data analysis result is a preset data analysis result, and output the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received. Namely, the invention supports data query, and sends the corresponding data analysis associated information to the user during query, so as to improve the user experience.
The previous description of the disclosed embodiments is provided to enable any person skilled in the art to make or use the present invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other embodiments without departing from the spirit or scope of the invention. Thus, the present invention is not intended to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.
Claims (5)
1. A method of data analysis, comprising:
acquiring initial heart rhythm data of a target user, and performing preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data; performing preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data, wherein the preset data processing operation comprises the following steps: performing data cleaning, data sorting and data downsampling operation on the initial heart rhythm data to obtain heart rhythm data to be processed; taking the heart rhythm data to be processed as target heart rhythm data;
calling a preset data analysis model to perform data processing on the target heart rhythm data so as to obtain a data analysis result corresponding to the target heart rhythm data; the preset data analysis model is obtained based on training data training; the training data comprises a heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample; the data analysis result is used for representing emotional states;
acquiring data analysis associated information under the condition that the data analysis result is a preset data analysis result, and outputting the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received;
the generation process of the preset data analysis model comprises the following steps: acquiring initial data samples arranged in a long-time sequence, and performing data cleaning, data sorting and data down-sampling operations on the initial data samples to obtain rhythm data samples; the initial data samples arranged in a long-time sequence are continuous time sequence heartbeat signals which exceed 24 hours; the data sorting comprises rearranging data according to a sequence from 0:00 hours to 24:00 hours;
acquiring a data analysis result label of the initial data sample, and using the data analysis result label as the data analysis result label of the heart rhythm data sample; training a preset data analysis model by using the heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample until a preset training stop condition is met, wherein the preset data analysis model is a time sequence random forest classification model;
after a preset data analysis model is called to perform data processing on the target rhythm data to obtain a data analysis result corresponding to the target rhythm data, the method further comprises the following steps:
acquiring data analysis results of the continuous preset number of the target users;
determining the number of data analysis results as preset data analysis results, and calculating the ratio of the number to the total number of the data analysis results;
and taking a data analysis result corresponding to the preset ratio interval in which the ratio is positioned as a data analysis correction result of the target user.
2. The data analysis method of claim 1, wherein obtaining initial heart rhythm data for a target user comprises:
acquiring an initial heart rhythm data set stored in a preset storage position; the initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data;
and acquiring target user information of a target user, and screening initial rhythm data corresponding to the target user information from the initial rhythm data set according to the user information corresponding to the initial rhythm data.
3. A data analysis apparatus, comprising:
the data processing module is used for acquiring initial heart rhythm data of a target user and carrying out preset data processing operation on the initial heart rhythm data to obtain target heart rhythm data;
the data analysis module is used for calling a preset data analysis model to perform data processing on the target heart rhythm data so as to obtain a data analysis result corresponding to the target heart rhythm data; the preset data analysis model is obtained based on training data; the training data comprises rhythm of heart data samples and data analysis result labels corresponding to the rhythm of heart data samples; the data analysis result is used for representing emotional states;
the data output module is used for acquiring data analysis associated information under the condition that the data analysis result is a preset data analysis result, and outputting the data analysis result and the data analysis associated information to the terminal under the condition that a data query request sent by the terminal of the target user is received;
the data analysis module is further used for obtaining the data analysis results of the continuous preset number of the target users after calling a preset data analysis model to perform data processing on the target heart rate data so as to obtain the data analysis results corresponding to the target heart rate data;
determining the number of data analysis results as preset data analysis results, and calculating the ratio of the number to the total number of the data analysis results;
taking a data analysis result corresponding to a preset ratio interval in which the ratio is positioned as a data analysis correction result of the target user;
the device also comprises a generation module of a preset data analysis model, a data sorting module and a data down-sampling module, wherein the generation module is used for acquiring initial data samples arranged in a long-term time sequence, and performing data cleaning, data sorting and data down-sampling operations on the initial data samples to obtain heart rhythm data samples; the initial data samples arranged in a long-time sequence are continuous time sequence heartbeat signals which exceed 24 hours; the data sorting comprises rearranging data according to a sequence from 0:00 hours to 24:00 hours; acquiring a data analysis result label of the initial data sample, and using the data analysis result label as a data analysis result label of the heart rhythm data sample; training a preset data analysis model by using the heart rhythm data sample and a data analysis result label corresponding to the heart rhythm data sample until a preset training stop condition is met, wherein the preset data analysis model is a time sequence random forest classification model;
the data processing module comprises: the data processing submodule is used for carrying out data cleaning, data sorting and data down-sampling operations on the initial heart rhythm data to obtain heart rhythm data to be processed; and taking the heart rhythm data to be processed as target heart rhythm data.
4. The data analysis device of claim 3, wherein the data processing module comprises:
the set acquisition submodule is used for acquiring an initial heart rhythm data set stored in a preset storage position; the initial heart rate data set comprises initial heart rate data acquired and uploaded by intelligent wearable devices worn by a plurality of users and user information corresponding to the initial heart rate data;
and the data screening submodule is used for acquiring target user information of a target user and screening initial heart rhythm data corresponding to the target user information from the initial heart rhythm data set according to the user information corresponding to the initial heart rhythm data.
5. An electronic device, comprising: a memory and a processor;
wherein the memory is used for storing programs;
a processor calls a program and is arranged to perform the data analysis method of any of claims 1-2.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210196265.8A CN114334090B (en) | 2022-03-02 | 2022-03-02 | Data analysis method and device and electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210196265.8A CN114334090B (en) | 2022-03-02 | 2022-03-02 | Data analysis method and device and electronic equipment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114334090A CN114334090A (en) | 2022-04-12 |
| CN114334090B true CN114334090B (en) | 2022-07-12 |
Family
ID=81030181
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210196265.8A Active CN114334090B (en) | 2022-03-02 | 2022-03-02 | Data analysis method and device and electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114334090B (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106419936A (en) * | 2016-09-06 | 2017-02-22 | 深圳欧德蒙科技有限公司 | Emotion classifying method and device based on pulse wave time series analysis |
| TW201801037A (en) * | 2016-06-30 | 2018-01-01 | 泰金寶電通股份有限公司 | Emotion analysis method and electronic apparatus thereof |
| CN109460752A (en) * | 2019-01-10 | 2019-03-12 | 广东乐心医疗电子股份有限公司 | Emotion analysis method and device, electronic equipment and storage medium |
| CN111067512A (en) * | 2019-12-05 | 2020-04-28 | 东软集团股份有限公司 | Ventricular fibrillation detection device, ventricular fibrillation detection model training method and equipment |
| CN112614583A (en) * | 2020-11-25 | 2021-04-06 | 平安医疗健康管理股份有限公司 | Depression grade testing system |
Family Cites Families (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109620262B (en) * | 2018-12-12 | 2020-12-22 | 华南理工大学 | Emotion recognition system and method based on wearable bracelet |
| CN109659009B (en) * | 2018-12-26 | 2022-05-31 | 杭州行为科技有限公司 | Emotion management method and device and electronic equipment |
| CN110705584A (en) * | 2019-08-21 | 2020-01-17 | 深圳壹账通智能科技有限公司 | Emotion recognition method, emotion recognition device, computer device and storage medium |
| EP3819386A1 (en) * | 2019-11-08 | 2021-05-12 | Alcediag | Diagnosing of mood disorders using blood rna editing biomarkers |
| CN112043252B (en) * | 2020-10-10 | 2021-09-28 | 山东大学 | Emotion recognition system and method based on respiratory component in pulse signal |
-
2022
- 2022-03-02 CN CN202210196265.8A patent/CN114334090B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| TW201801037A (en) * | 2016-06-30 | 2018-01-01 | 泰金寶電通股份有限公司 | Emotion analysis method and electronic apparatus thereof |
| CN106419936A (en) * | 2016-09-06 | 2017-02-22 | 深圳欧德蒙科技有限公司 | Emotion classifying method and device based on pulse wave time series analysis |
| CN109460752A (en) * | 2019-01-10 | 2019-03-12 | 广东乐心医疗电子股份有限公司 | Emotion analysis method and device, electronic equipment and storage medium |
| CN111067512A (en) * | 2019-12-05 | 2020-04-28 | 东软集团股份有限公司 | Ventricular fibrillation detection device, ventricular fibrillation detection model training method and equipment |
| CN112614583A (en) * | 2020-11-25 | 2021-04-06 | 平安医疗健康管理股份有限公司 | Depression grade testing system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114334090A (en) | 2022-04-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Liang et al. | A new, short-recorded photoplethysmogram dataset for blood pressure monitoring in China | |
| Speer et al. | Measuring heart rate variability using commercially available devices in healthy children: A validity and reliability study | |
| WO2018117914A1 (en) | Determining wellness using activity data | |
| Li et al. | Eustress or distress: An empirical study of perceived stress in everyday college life | |
| JP7240789B2 (en) | Systems for screening and monitoring of encephalopathy/delirium | |
| CN108778097A (en) | Device and method for assessing heart failure | |
| KR20090024808A (en) | Assessing dementia and dementia-type disorders | |
| Mayya et al. | Continuous monitoring of stress on smartphone using heart rate variability | |
| Khowaja et al. | Toward soft real-time stress detection using wrist-worn devices for human workspaces | |
| CN113133752B (en) | Psychological assessment method, system, device and medium based on heart rate variability analysis | |
| Lamichhane et al. | Towards stress detection in real-life scenarios using wearable sensors: normalization factor to reduce variability in stress physiology | |
| WO2022095331A1 (en) | Stress assessment method and apparatus, computer device and storage medium | |
| Majumder et al. | sEmoD: A personalized emotion detection using a smart holistic embedded IoT system | |
| Fu et al. | Symmetric convolutional and adversarial neural network enables improved mental stress classification from EEG | |
| CN115089179A (en) | Psychological emotion insights analysis method and system | |
| Chou et al. | Effects of ecg data length on heart rate variability among young healthy adults | |
| CN113764105A (en) | Cardiovascular data prediction method for middle-aged and old people | |
| Polak et al. | Processing photoplethysmograms recorded by smartwatches to improve the quality of derived pulse rate variability | |
| CN114334090B (en) | Data analysis method and device and electronic equipment | |
| CN115966308A (en) | Method, system and device for detecting and quantifying sleep memory activity level | |
| Park et al. | A study on the development of a day-to-day mental stress monitoring system using personal physiological data | |
| CN110192872B (en) | Pressure evaluation calibration method and device and storage medium | |
| Ni et al. | Extracting heartbeat intervals using self-adaptive method based on ballistocardiography (BCG) | |
| McLean et al. | A Sliding Scale Signal Quality Metric of Photoplethysmography Applicable to Measuring Heart Rate across Clinical Contexts with Chest Mounting as a Case Study | |
| Bouarfa et al. | Profiling visual and verbal stress responses using electrodermal heart rate and hormonal measures |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |