Subspace KNN Research Articles

Detection of cardiac arrhythmias from ECG signals using FBSE and Jaya optimized ensemble random subspace K-nearest neighbor algorithm

Cardiac Arrhythmias is defined as the sudden and unexpected occurrence of a rhythm in the heart. Arrhythmia detection is prevalent in developing countries that require systems-based screening solutions for diagnosis. A new rhythm-based approach is proposed to screen the patients with cardiac arrhythmia at the primary level. This method eliminates multiple tests, enabling faster and more accurate diagnosis. This method first segments the electrocardiogram signals to create a complete picture of the single heartbeat. Then the Fourier–Bessel series expansion (FBSE) is used to transform the sequences of each heartbeat into more meaningful ones that can characterize the structural integrity of arrhythmia. The FBSE sequence series are trained using the Jaya optimized ensemble random subspace KNN (JO-ERSKNN) model with 10-fold cross-validation for classifying five types of cardiac arrhythmia beats. We have achieved an accuracy of 99.49%, a sensitivity of 95.43%, and specificity of 99.48%. The results demonstrate that the proposed algorithm can detect differences in the five types of cardiac arrhythmia signals. It can also be utilized as a screening tool for detecting arrhythmia and can be made compatible with various wearable devices.

Surface and Underwater Acoustic Source Discrimination Based on Machine Learning Using a Single Hydrophone

In shallow water, passive sonar usually has great difficulty in discriminating a surface acoustic source from an underwater one. To solve this problem, a supervised machine learning method using only one hydrophone is implemented in this paper. Firstly, simulated training data are generated by a normal mode model KRAKEN with the same environment setup as that in SACLANT 1993 experiment. Secondly, the k-nearest neighbor (kNN) classifiers are trained and evaluated using the scores of precision, recall, F1 and accuracy. Thirdly, the random subspace kNN classifiers are finely trained on three hyperparameters (the number of nearest neighbors, the number of predictors selected at random and the number of learners in the ensemble) to obtain the best model. Fourthly, a deep learning method called ResNet-18 is also applied, and it reaches the best balance between precision and recall, while the accuracies of both simulation and experimental data are all 1.0. Further, data from all 48 hydrophones of the vertical linear array (VLA) are analyzed using the three kinds of machine learning methods (kNN, random subspace kNN and ResNet-18) separately, and the results are compared. It is concluded that the performance of random subspace kNN is the best. Both the simulation and experimental results suggest the feasibility of machine learning as a surface and underwater acoustic source discrimination method even with only a single hydrophone.

Machine Learning-Based Fault Diagnosis for a PWR Nuclear Power Plant

Fault detection and diagnosis (FDD) systems can reduce high costs and energy consumption. This paper presents a machine learning-based fault detection and diagnosis (FDD) technique for actuators and sensors in a pressurized water reactor (PWR). In the proposed FDD framework, faults are first detected using a shallow neural network. Second, fault diagnosis is performed using 15 different classifiers provided in the MATLAB Classification Learner toolbox, including support vector machine (SVM), K-nearest neighbor (KNN), and ensemble. Several classifiers were found to provide superior classification performance, including medium KNN, cubic KNN, cosine KNN, weighted KNN, fine Gaussian SVM, quadratic SVM, medium Gaussian SVM, coarse Gaussian, bagged trees, and subspace KNN. The accuracy of the FDD approach was demonstrated using a set of simulation results.

Geographical origin determination of Red Huajiao in China using the electronic nose combined with ensemble recognition algorithm.

Red Huajiao was the most important Zanthoxylum species in China, and its quality was highly determined the geographical region. This study was aimed to establish a determination method for the geographical origin recognition of Red Huajiao by using the electronic nose and ensemble recognition algorithm. Six origins of samples were detected by the electronic nose, and two categories of electronic nose sensors characteristic values, named as "optimized characteristic value" and "filtered characteristic value," were obtained by the principal component analysis and discrimination index method and Filter-Wrapper method. Based on the two categories of characteristic values, 22 kinds of model analysis methods, which belonged to five categories of ensemble recognition algorithms were used to recognize the geographical origin. The total recognition accuracy rate of the two categories of characteristic values were 83.9% and 85.7%, respectively. Furthermore, during 22 kinds of model analysis method, the ensemble Subspace KNN and Bagged Trees methods in Ensemble Learning algorithm exhibited the best distinguishing ability with the accuracy rate more than 90%. Therefore, the electronic nose combined with Ensemble Learning would be promising for the geographical origin determination application. PRACTICAL APPLICATION: This work demonstrates that the Red Huajiao can be simply and rapidly determined by using electronic nose combined with ensemble recognition algorithm, allowing to effectively distinguish geographical origin of Red Huajiao, which can provide an important reference for the quality assessment of Huajiao.

COMPARATIVE ANALYSIS OF MACHINE LEARNING ALGORITHMS FOR THE DETECTION OF ANDROID MALWARE

This paper examines the effectiveness of some machine learning algorithms in the detection of android malicious application. In order to carry out this analysis, drebin dataset of android malicious and good applications were obtained and used for the classification as described in a section of this article. The classification results show that the Cubic SVM, Quadratic SVM and ensemble Subspace KNN performed better with 99.2%, 98.7% and 98.4% accuracy with 0.0079, 0.0129 and 0.1598 error rate respectively.

Prediction of student attrition risk using machine learning

PurposeThe prediction of student attrition is critical to facilitate retention mechanisms. This study aims to focus on implementing a method to predict student attrition in the upper years of a physiotherapy program.Design/methodology/approachMachine learning is a computer tool that can recognize patterns and generate predictive models. Using a quantitative research methodology, a database of 336 university students in their upper-year courses was accessed. The participant's data were collected from the Financial Academic Management and Administration System and a platform of Universidad Autónoma de Chile. Five quantitative and 11 qualitative variables were chosen, associated with university student attrition. With this database, 23 classifiers were tested based on supervised machine learning.FindingsAbout 23.58% of males and 17.39% of females were among the attrition student group. The mean accuracy of the classifiers increased based on the number of variables used for the training. The best accuracy level was obtained using the “Subspace KNN” algorithm (86.3%). The classifier “RUSboosted trees” yielded the lowest number of false negatives and the higher sensitivity of the algorithms used (78%) as well as a specificity of 86%.Practical implicationsThis predictive method identifies attrition students in the university program and could be used to improve student retention in higher grades.Originality/valueThe study has developed a novel predictive model of student attrition from upper-year courses, useful for unbalanced databases with a lower number of attrition students.

An accurate automated schizophrenia detection using TQWT and statistical moment based feature extraction

Nowadays, abnormal brain activities can be automatically detected and classified by processing EEG signals. In this paper, the classification process of EEG signals collected from healthy and schizophrenic individuals was carried out. For this purpose, firstly, feature extraction from 19-channel EEG signals with healthy and schizophrenia classes was performed using Tunable Q-Factor Wavelet Transform (TQWT) and statistical moment methods. After this process, feature selection was made using the ReliefF method and the extracted features were reduced. In the last stage of the study, these features were classified using the ensemble subspace k-Nearest Neighbor (kNN) method. As a result of the proposed method, performance values ​​for each channel were calculated and 99.12 % accuracy was achieved. When the obtained results are compared with the studies in the literature, it is seen that the method is extremely fast and successful.

Minimum Mapping from EMG Signals at Human Elbow and Shoulder Movements into Two DoF Upper-Limb Robot with Machine Learning

This research focuses on the minimum process of classifying three upper arm movements (elbow extension, shoulder extension, combined shoulder and elbow extension) of humans with three electromyography (EMG) signals, to control a 2-degrees of freedom (DoF) robotic arm. The proposed minimum process consists of four parts: time divisions of data, Teager–Kaiser energy operator (TKEO), the conventional EMG feature extraction (i.e., the mean absolute value (MAV), zero crossings (ZC), slope-sign changes (SSC), and waveform length (WL)), and eight major machine learning models (i.e., decision tree (medium), decision tree (fine), k-Nearest Neighbor (KNN) (weighted KNN, KNN (fine), Support Vector Machine (SVM) (cubic and fine Gaussian SVM), Ensemble (bagged trees and subspace KNN). Then, we compare and investigate 48 classification models (i.e., 47 models are proposed, and 1 model is the conventional) based on five healthy subjects. The results showed that all the classification models achieved accuracies ranging between 74–98%, and the processing speed is below 40 ms and indicated acceptable controller delay for robotic arm control. Moreover, we confirmed that the classification model with no time division, with TKEO, and with ensemble (subspace KNN) had the best performance in accuracy rates at 96.67, recall rates at 99.66, and precision rates at 96.99. In short, the combination of the proposed TKEO and ensemble (subspace KNN) plays an important role to achieve the EMG classification.

Classification of Epileptic EEG Signals Using Synchrosqueezing Transform and Machine Learning.

Epilepsy is a neurological disease that is very common worldwide. Patient's electroencephalography (EEG) signals are frequently used for the detection of epileptic seizure segments. In this paper, a high-resolution time-frequency (TF) representation called Synchrosqueezing Transform (SST) is used to detect epileptic seizures. Two different EEG data sets, the IKCU data set we collected, and the publicly available CHB-MIT data set are analyzed to test the performance of the proposed model in seizure detection. The SST representations of seizure and nonseizure (pre-seizure or inter-seizure) EEG segments of epilepsy patients are calculated. Various features like higher-order joint TF (HOJ-TF) moments and gray-level co-occurrence matrix (GLCM)-based features are calculated using the SST representation. By using single and ensemble machine learning methods such as k-Nearest Neighbor (kNN), Logistic Regression (LR), Naive Bayes (NB), Support Vector Machine (SVM), Boosted Trees (BT), and Subspace kNN (S-kNN), EEG features are classified. The proposed SST-based approach achieved 95.1% ACC, 96.87% PRE, 95.54% REC values for the IKCU data set, and 95.13% ACC, 93.37% PRE, 90.30% REC values for the CHB-MIT data set in seizure detection. Results show that the proposed SST-based method utilizing novel TF features outperforms the short-time Fourier transform (STFT)-based approach, providing over 95% accuracy for most cases, and compares well with the existing methods.

Tactile Sensing Using Machine Learning-Driven Electrical Impedance Tomography

Electrical Impedance Tomography (EIT) tactile sensors have limited success in equipping robots with tactile sensing capabilities due to the low spatial resolution of the resulting tactile images and the presence of image artifacts. To address these limitations, we propose a modular framework for invariant recognition of objects, within the context of an EIT artificial skin sensor. Three interconnected problems, i.e., EIT image reconstruction, segmentation and object recognition, are tackled in this work with the aid of machine learning. A novel conductivity surface decomposition approach, based on low order bivariate polynomials and RBF networks is introduced for the efficient solution of the EIT inverse problem. Next, segmentation of the reconstructed images is performed using a convolutional neural network and transfer learning. Finally, a subspace KNN ensemble classifier is trained on the set of object descriptors extracted from the segmented inhomogeneities to classify the objects. The proposed framework provides an accuracy of 97.5% on unseen data.

Optimal EEG channels selection for alcoholism screening using EMD domain statistical features and harmony search algorithm

Alcoholism can be analyzed by Electroencephalogram (EEG) data. Finding an optimal subset of EEG channels for alcoholism detection is a challenging task. The paper reports a new methodology for the detection of optimal channels for alcoholism analysis using EEG data. The proposed technique employs the Empirical Mode Decomposition (EMD) technique to extract the amplitude and frequency modulated bandwidth features from the Intrinsic Mode Function (IMF) and ensemble subspace K-NN as a classifier to classify alcoholics and normal. The optimum channels are selected, using a harmony search algorithm. The fitness value of discrete binary harmony search (DBHS) optimization algorithms is calculated using accuracy and sensitivity achieved by the ensemble subspace K-Nearest Neighbor classifier. Experimental outcomes indicate that the optimal channel selected by the harmony search algorithm has biological inference related to the alcoholic subject. The proposed approach reports a classification accuracy of 93.87%, with only 12 detected EEG channels.

Saliency-based classification of objects in unconstrained underwater environments

Exploration of the deep-sea underwater environment is a challenging and non-trivial task. Underwater vehicles used for the exploration of such environments capture videos continuously. The processing of these videos is a major bottleneck for scientific research in this area. This paper presents a methodology for the classification of the objects in the unconstrained underwater environments into two broad classes namely - man-made and natural. The classification of the objects is achieved using the saliency gradient based morphological active contour models. A bag of features acquired from the contours of the objects is used for the classification using various classifiers. Principal Component Analysis is used for the removal of redundancy in the feature set. The proposed features classify the objects present in the unconstrained underwater environment into a man-made/natural class using the proposed features. The results show that all the classifiers performed well; though KNN and ensemble subspace KNN, performed marginally better.

Real time terrain identification of autonomous robots using machine learning

In this project, machine learning based techniques for real time terrain identification of the autonomous robots are investigated. The factors affecting the performance of autonomous robots include nature of trajectories, on-course obstacles, and nature of terrain. The challenges involved in understanding the terrain of autonomous robots are called localization problems. This project investigates a robust classification based machine learning model to identify the terrains of an autonomous robot from a set of input sensor data , which would incorporated as features in the model. The features are selected with respect to the kinematic and dynamic model of differential drive robots. The terrains are classified into 11 classes and the inputs from different sensors are measured and categorized into the respective classes. A total of 49345 readings were taken. Twenty three classification learning methods are evaluated to find the best fitting model that can identify the terrains of robots in real time. Ensemble Subspace KNN classification learning model produced an accuracy of 100 %, observed as the best model for terrain identification. The results are represented using confusion matrix, which shows the relation between original terrains and model predicted terrains , scatter plot that represents the relationship between each features and ROC Curve analyses each sensor input data. The model output can be provided to an intelligent mechanism to control the wheels of robots and improve their performance.

Classification of visual cognitive workload using analytic wavelet transform

Visual stimuli are used to induce cognitive load of various levels to examine its effect on EEG signals. Four cognitive workload levels are generated by using different shapes and their arrangements in the visual field. Assessing the cognitive load imposed by visual demand is required for developing viable visual interfaces. In this experiment, EEG signals are acquired from 44 subjects by displaying them various cognitive level visual stimuli. The features for the same cognitive level vary significantly across subjects. This poses a complex problem for the classification of cognitive workload levels. The features for classification are obtained by decomposing EEG signals using the analytic wavelet transform. This transform allows to vary Q factor to improve high-frequency resolution. Kurtosis is used as a statistical feature. To capture local variations of the signal local binary pattern is used. The other features computed include morphological features such as curve length and non-linear energy. By using Ensemble subspace k-nearest neighbor and support vector machine, we have done 10-fold cross-validation for obtaining results. We have obtained comparable performance for similar studies. EEG channels are ranked based on weight obtained using neighborhood component feature selection method. The impact of the EEG channel on the detection of cognitive workload level is investigated.

Diagnosis of breast cancer with Stacked autoencoder and Subspace kNN

Breast cancer is one of the most common and deadliest cancer types in women worldwide. Research on this disease has become very important because early diagnosis stages, clinical applications and the speed of response to treatment are facilitated in diseases such as cancer. In this study, an approach is proposed in which a Subspace kNN algorithm is used together with Stacked autoencoder (SAE) for diagnosis of disease on the breast cancer microarray dataset for the first time. Such hybrid approaches can provide better results when classifying data sets with high-dimensional and uncertainty. The data set used in the study was taken from Kent Ridge-2 database. It consists of 97 samples (51 benign, 46 malicious) and 24482 attributes. The performance of the proposed method was evaluated and the results were compared with other well-known methods of dimension reduction and machine learning. As a result of the comparison, the data set was reduced to 100 attributes by using SAE and Subspace kNN and 91.24% accuracy was achieved. The result obtained provides important classification accuracy, especially in high-dimensional data sets. The importance of this study is that the models that were created by using various classifiers to increase the success rate of the stacked autoencoder-softmax classifier model in the breast cancer microarray data set were applied for the first time. In this regard, it is considered that automation-based studies will provide diagnostic decision support system a solution using the proposed method in future works.

A flexible analytic wavelet transform based approach for motor-imagery tasks classification in BCI applications

Background and Objective: Motor Imagery (MI) based Brain-Computer-Interface (BCI) is a rising support system that can assist disabled people to communicate with the real world, without any external help. It serves as an alternative communication channel between the user and computer. Electroencephalogram (EEG) recordings prove to be an appropriate choice for imaging MI tasks in a BCI system as it provides a non-invasive way for completing the task. The reliability of a BCI system confides on the efficiency of the assessment of different MI tasks.Methods: The present work proposes a new approach for the classification of distinct MI tasks based on EEG signals using the flexible analytic wavelet transform (FAWT) technique. The FAWT decomposes the EEG signal into sub-bands and temporal moment-based features are extracted from the sub-bands. Feature normalization is applied to minimize the bias nature of classifier. The FAWT-based features are utilized as inputs to multiple classifiers. Ensemble learning method based Subspace k-Nearest Neighbour (kNN) classifier is established as the best and robust classifier for the distinction of the right hand (RH) and right foot (RF) MI tasks.Results: The sub-band (SB) wise features are tested on multiple classifiers and best performance parameters are obtained using the ensemble method based subspace kNN classifier. The best results of parameters are obtained for fourth SB as accuracy 99.33%, sensitivity 99%, specificity 99.6%, F1-Score 0.9925, and kappa value 0.9865. The other sub-bands are also attained significant results using subspace KNN classifier.Conclusions: The proposed work explores the utility of FAWT based features for the classification of RH and RF MI tasks EEG signals. The suggested work highlights the effectiveness of multiple classifiers for classification MI-tasks. The proposed method shows better performance in comparison to state-of-arts methods. Thus, the potential to implement a BCI system for controlling wheelchairs, robotic arms, etc.

Classification of heart sounds using fractional fourier transform based mel-frequency spectral coefficients and traditional classifiers

Heart sounds are a rich source of information for early diagnosis of cardiac pathologies. Distinguishing normal from abnormal heart sounds requires a specially trained clinician. Our goal is to develop a machine learning application that tackle the problem of heart sound classification. So we present a new processing and classification system for heart sounds. The automated diagnostic system is described in terms of its preprocessing, cardiac cycle segmentation, feature extraction, features reduction and classification stages. Conventional architectures will be used to identify abnormal heart sounds then the performances of the proposed systems will be compared. The conventional architectures include the following traditional classifiers: SVM, KNN and ensemble classifier (bagged Trees, subspace KNN and RUSBoosted tree). The proposed system is verified on the publicly available dataset of the heart sounds. The cross-validation and local hold out train-test methods are used to perform the experiments and obtain and compare the results. The proposed system showed potential for achieving excellent performance compared to previous methods on the same dataset with a score of 0.9200 at a sensitivity of 0.8735 and specificity of 0.9666 using a support vector machine classifier with cubic kernel. The details of the methodology and the results are presented and discussed.

THE EFFICIENCY OF ENSEMBLE CLASSIFIERS IN PREDICTING THE JOHANNESBURG STOCK EXCHANGE ALL-SHARE INDEX DIRECTION

The success of investors in obtaining huge financial rewards from the stock market depends on their ability to predict the direction of the stock market index. The purpose of this study is to evaluate the efficacy of several ensemble prediction models (Boosted, RUS-Boosted, Subspace Disc, Bagged, and Subspace KNN) in predicting the daily direction of the Johannesburg Stock Exchange (JSE) All-Share index compared to other commonly used machine learning techniques including support vector machines (SVM), logistic regression and [Formula: see text]-nearest neighbor (KNN). The findings in this study show that, among all ensemble models, Boosted algorithm is the best performer followed by RUS-Boosted. When compared to the other techniques, ensemble technique (represented by Boosted) outperformed these techniques, followed by KNN, logistic regression and SVM, respectively. These findings suggest that investors should include ensemble models among the index prediction models if they want to make huge profits in the stock markets. However, not all investors can benefit from this as models may suffer from alpha decay as more and more investors use them, implying that the successful algorithms have limited shelf life.

Bleeding Detection in Wireless Capsule Endoscopy Image Video Using Superpixel-Color Histogram and a Subspace KNN Classifier.

Wireless Capsule Endoscopy (WCE) has become increasingly popular in clinical gastrointestinal (GI) disease diagnosis, benefiting from its painless and noninvasive examination. However, reviewing a large number of images is time-consuming for doctors, thus a computer-aided diagnosis (CAD) system is in high demand. In this paper, we present an automatic bleeding detection algorithm that consists of three stages. The first stage is the preprocessing, including key frame extraction and edge removal. In the second stage, we discriminate the bleeding frames using a novel superpixelcolor histogram (SPCH) feature based on the principle color spectrum, and then the decision is made by a subspace KNN classifier. Thirdly, we further segment the bleeding regions by extracting a 9-D color feature vector from the multiple color spaces at the superpixel level. Experimental results with an accuracy of 0.9922 illustrate that our proposed method outperforms the state-of-the-art methods in GI bleeding detection with low computational costs.

Noncontact Sleep Stage Estimation Using a CW Doppler Radar

Sleep stage estimation is crucial to the evaluation of sleep quality and is a proven biometric in diagnosing cardiovascular diseases. In this paper, we design a continuous wave (CW) Doppler radar to accurately measure sleep-related signals, including respiration, heartbeat, and body movement. Body movement index, respiration per minute (RPM), variance of RPM, amplitude difference accumulation (ADA) of respiration, rapid eye movement parameter, sample entropy, heartbeat per minute (HPM), variance of HPM, ADA of heartbeat, deep parameter, and time feature have been extracted and fed into different machine learning classifiers. A total of 11 all night polysomnography recordings from 13 healthy examinees were used to validate the proposed CW Doppler radar system and the ability to detect sleep stage information from it. Comparative studies and statistical results have shown that the subspace K-nearest neighbor algorithm outperforms the other classifiers with the highest accuracy of up to 86.6%. With the Relief F algorithm, features have been ranked, and the selected feature subsets have been preliminary tested to identify the optimal feature subset. Meanwhile, comparative analysis of our classification performance under different sleep stage patterns with prior works has been carried out to show the significant improvements over state-of-the-art solutions. These results suggest that the proposed scheme is suitable for long-term sleep monitoring.