Design And Implementation Of Emotional Analysis And Social Sharing System Based On ECG Data

In recent years,with the rapid development of science and technology,more and more attention has been paid to the research on the emotional interaction between humans and computers,which makes emotional computing a hot research topic.Emotion recognition is the most important part of emotional calculation,which mainly involves body posture,speech signal,facial expression and physiological signals.Compared with other aspects,physiological signals which are not influenced by subjective factors are more authentic and objective,so the emotion recognition based on physiological signals is more profound.As one of the most important physiological signals of the human body,ECG signals can objectively reflect the changes in human emotions.Therefore,ECG signals are used for emotional recognition research.The specific research process is divided into three steps:the preconditioning of ECG signals,the extraction of emotional eigenvalues and emotion recognition of eigenvalues.This article selects 511 segments with good ECG waveforms,including calm,pleasure,sadness,anger,disgust,fear,and surprise,as the experimental data.Wavelet threshold denoising is used to preprocess these data to remove noise and baseline drift.Continuous wavelet transform method was used to detect the PQRST key points of the ECG waveform.Then,using the positions of P,Q,R,S,and T key points as parameters,a total of ECG signals was extracted in the time and frequency domain.78 statistical eigenvalues.For the classification of emotional categories of seven types of emotional data,this paper first analyzes and studies the following types of sentiment recognition methods based on ECG data:a method based on maximum-minimum ant colony algorithm(MMAS)combined with KNN classification algorithm,the method based on BP neural network,the method based on convolutional neural network(CNN),one-to-one support vector machine(SVM)classification algorithm,SVM classification algorithm based on complete binary tree,one-to-one linear discriminant analysis(LDA)classification algorithm and LDA classification algorithm based on complete binary tree.Then use these seven emotion recognition methods to carry out emotion recognition experiments on seven types of emotion sample data.Experimental results show that the emotion recognition method based on LDA classifier has the best classification effect on the seven emotions of ECG data.Therefore,using this emotion recognition method as the basis of the algorithm,this paper develops and implements a sentiment analysis and social sharing system based on ECG data in Android Studio environment using Bluetooth development technology,data mapping technology and JNI programming technology.The system includes five major functional modules:electrocardiogram display function module,sentiment analysis function module,social sharing function module,basic setup function module and other functional modules.The system can feedback the emotional state of the mentally ill patients to the doctor anytime and anywhere to help the psychiatrist to understand and adjust the patient's emotional state in a timely manner so that the emotions of the mentally ill patients can develop toward an active and healthy direction so as to achieve the purpose of mobile medical treatment.