PCA For Feature Extraction Research Articles

Automatic Heart Sound Segmentation and Murmur Detection in Phonocardiogram

This article aimed to design a software system for partial heart sound and define the problem to facilitate the doctor's primary job of diagnosing the normal or abnormal voice. The dataset was taken from Kaggle online site. The system was designed on several steps. The first step was to remove the noise by the use of a second-order bandpass filter, and the cutoff frequency was close to the heart sound frequency. The second step was to remove the spike by spike removal, and the final step was to use downsampling to normalize the amplitude of the signal to facilitate using the artificial neural network. The step that followed the signal processing was the segmentation and was done using Hilbert envelop, and then wavelet and PCA for feature extraction were used; finally, classification and training were done by feed forward neural network. All of these steps were implanted in MATLAB, and tests have shown that the accuracy is 97.639%. The number of data used in the training was 138 signal.

Classification of Human Emotion Using DT-SVM Algorithm with Enhanced Feature Selection and Extraction

Emotions are a basic component of human life. It generates different brain waves for emotions such as happiness, sadness, anger, calmness, tension, excitement, etc. The brain waves are electric and their electric impulse can be measured and recorded as a continuous stream of data. These emitted brain waves are recorded using an EEG device. Many existing systems are in use that feeds the recorded data into various Machine learning algorithms to classify the emotions. These systems are huge and complex, thus require a great amount of time for initializing and working. While a lot of algorithms are used and new algorithms are discovered to classify Brain EEG data, most of the time results will be improper and will not be reliable. The proposed system extracts only the data which corresponds to Human-emotions from the continuous stream of EEG data. The system makes use of robust preprocessing algorithms like ANOVA and PCA for feature extraction and selection to identify and extract features associated with Human-emotion. Later, these recording signals are modeled and fed into Dynamic Time wrapping Simple vector machine (DT-SVM) classification algorithm to analyze and predict the emotion of the person during the experiment which produces an improved accuracy of 99.2%compared to existing system.

PCA-Based on Feature Extraction and Compressed Sensing for Dimensionality Reduction

Compressive sensing reduces the number of samples required to achieve acceptable reconstruction for medical diagnostics, therefore this research will implement dimensional reduction algorithms through compressed sensing for electrocardiogram signals (EKG). dimensional reduction is performed based on the fact that ECG signals can be reconstructed with linear combination coefficients with a bumpy base of small measurements with high accuracy. This study will use PCA for feature extraction on ECG signals. The data used are the ECG patient records on the website page www.physionet.org as many as 1200 with each attribute as many as 256 attributes. The total data dimension used is 1200x256, which means the data has 1200 rows and has as many as 256 columns. To show the accuracy of the dimensional reduction result, so it is performed classification on data using KNN and Naive Bayes. The classification results show that KKN can classify well with 84,02% accuracy rate and the Naive Bayes accuracy is 65,78%. for 100 dimensions The conclusion is those dimensional reductions for ECG data that have large dimensions, it still able to provide valid information like it uses the original data. Principle Component Analysis is a good method for reducing data dimensions by selecting certain features, so the dimensions of the data become smaller but still able to provide good accuracy to the reader.

Gear fault feature extraction and classification of singular value decomposition based on Hilbert empirical wavelet transform

Vibration signal of gearbox systems carries the important dynamic information for fault diagnosis. However, vibration signals always show non stationary behavior and overwhelmed by a large amount of noise make this task challenging in many cases. Thus, a new fault diagnosis method combining the Hilbert empirical wavelet transform (HEWT), the singular value decomposition (SVD) and Elman neural network is proposed in this paper. Vibration signals of normal gear, gear with tooth root crack, gear with chipped tooth in width, gear with chipped tooth in length, gear with missing tooth and gear with general surface wear are collected in different speed and load conditions. HEWT, a new self-adaptive time-frequency analysis, was applied to the vibration signals to obtain the instantaneous amplitude matrices. Singular value vectors, as the fault feature vectors were then acquired by applying the SVD. Last, the Elman neural network was used for automatic gearbox fault identification and classification. Through experimental results, it was concluded that the proposed method can accurately extract and classify the gear fault features under variable conditions. Moreover, the performance of the proposed HEWT-SVD method has an advantage over that of Hilbert-Huang transform (HHT)-SVD, local mean decomposition (LMD)-SVD or wavelet packet transform (WPT)-PCA for feature extraction.

Data Driven Models for Satellite State-of-Health Monitoring and Evaluation

Satellite state-of-health conditions and keeping up systems safe are the fundamental worry of space operations. Since managing the correlation structure between variables give greater information than individual variables or causes relationships, Data driven methodology presented as an answer of managing enormous information. In this article a survey of some recent related work is done. Then a mathematical foundation for latent variable modeling, support vector machines (SVMs) and artificial neural networks (ANN) is illustrated. Then a four data driven models are introduced for monitoring, diagnosis and prediction of the thermal control of satellite power supply system. Two kinds of PLS-Batch models were implemented for the first time with a satellite telemetry. After that a utilization of multivariate Shewart, DModX in addition to CuSum charts for system monitoring and early fault detection. Another two machine learning models were built and coded in Python. These two models are (SVMs) model and (ANN) model. Both of the last two models are integrated with PCA for feature extraction and dimensionality reduction for visualization. The outcomes confirmed the effectiveness of utilizing such models in monitoring, faults detection, and diagnosis. Also system health evaluation could be established within the operation through these models. In addition to prediction of system behavior and early faults detection.