Feature Selection for Interpatient Supervised Heart Beat Classification

Supervised and interpatient classification of heart beats is primordial in many applications requiring long-term monitoring of the cardiac function. Several classification models able to cope with the strong class unbalance and a large variety of feature sets have been proposed for this task. In practice, over 200 features are often considered, and the features retained in the final model are either chosen using domain knowledge or an exhaustive search in the feature sets without evaluating the relevance of each individual feature included in the classifier. As a consequence, the results obtained by these models can be suboptimal and difficult to interpret. In this work, feature selection techniques are considered to extract optimal feature subsets for state-of-the-art ECG classification models. The performances are evaluated on real ambulatory recordings and compared to previously reported feature choices using the same models. Results indicate that a small number of individual features actually serve the classification and that better performances can be achieved by removing useless features.

Download Full-text

A Unified View of Causal and Non-causal Feature Selection

ACM Transactions on Knowledge Discovery from Data ◽

10.1145/3436891 ◽

2021 ◽

Vol 15 (4) ◽

pp. 1-46

Author(s):

Kui Yu ◽

Lin Liu ◽

Jiuyong Li

Keyword(s):

Feature Selection ◽

Bayesian Network ◽

Synthetic Data ◽

Selection Methods ◽

Bayesian Network Model ◽

Real World Data ◽

Feature Sets ◽

Unified View ◽

Optimal Feature ◽

Different Levels

In this article, we aim to develop a unified view of causal and non-causal feature selection methods. The unified view will fill in the gap in the research of the relation between the two types of methods. Based on the Bayesian network framework and information theory, we first show that causal and non-causal feature selection methods share the same objective. That is to find the Markov blanket of a class attribute, the theoretically optimal feature set for classification. We then examine the assumptions made by causal and non-causal feature selection methods when searching for the optimal feature set, and unify the assumptions by mapping them to the restrictions on the structure of the Bayesian network model of the studied problem. We further analyze in detail how the structural assumptions lead to the different levels of approximations employed by the methods in their search, which then result in the approximations in the feature sets found by the methods with respect to the optimal feature set. With the unified view, we can interpret the output of non-causal methods from a causal perspective and derive the error bounds of both types of methods. Finally, we present practical understanding of the relation between causal and non-causal methods using extensive experiments with synthetic data and various types of real-world data.

Download Full-text

CLASSIFICATION OF HIGH-DIMENSIONAL MICROARRAY DATA WITH A TWO-STEP PROCEDURE VIA A WILCOXON CRITERION AND MULTILAYER PERCEPTRON

International Journal of Computational Intelligence and Applications ◽

10.1142/s1469026811002969 ◽

2011 ◽

Vol 10 (01) ◽

pp. 1-14

Author(s):

VLADIMIR NIKULIN ◽

TIAN-HSIANG HUANG ◽

GEOFFREY J. MCLACHLAN

Keyword(s):

Data Mining ◽

Feature Selection ◽

High Dimensional ◽

Second Step ◽

Support Vector ◽

Step Procedure ◽

Leave One Out ◽

Natural Combination ◽

Feature Selection Techniques

The method presented in this paper is novel as a natural combination of two mutually dependent steps. Feature selection is a key element (first step) in our classification system, which was employed during the 2010 International RSCTC data mining (bioinformatics) Challenge. The second step may be implemented using any suitable classifier such as linear regression, support vector machine or neural networks. We conducted leave-one-out (LOO) experiments with several feature selection techniques and classifiers. Based on the LOO evaluations, we decided to use feature selection with the separation type Wilcoxon-based criterion for all final submissions. The method presented in this paper was tested successfully during the RSCTC data mining Challenge, where we achieved the top score in the Basic track.

Download Full-text

Automatic Classification of ECG Using Multi-Classifiers with a New Fusion Method

Applied Mechanics and Materials ◽

10.4028/www.scientific.net/amm.687-691.3917 ◽

2014 ◽

Vol 687-691 ◽

pp. 3917-3922

Author(s):

Yi Chang Wang ◽

Feng Qi Yan ◽

Yu Fang

Keyword(s):

Traditional Method ◽

Heart Beat ◽

Vector Model ◽

Support Vector ◽

Beat Classification ◽

Computer Aided Analysis ◽

Important Basis ◽

Voting Method ◽

Better Than

ECG signal contains abundant information of human heart activity. It is important basis of doctors’ diagnose. With the development of computer technology, computer aided analysis has been widely applied in the field of ECG analysis. Most of the traditional method is based on single classifier and too complex. Also, the accuracy is not high. This paper focuses on ECG heart beat classification, extracting different types of feature, training different classifiers by vector model and support vector machine (SVM), merging the result of multiple classifiers. In this paper, we used the advanced voting method (voting by weight) to fusion the result of different classifier, having compared it with the traditional voting method.It performed better than traditional method in term of accuracy

Download Full-text

Hybrid Ensemble Learning Methods for Classification of Microarray Data

Data Analytics in Medicine ◽

10.4018/978-1-7998-1204-3.ch038 ◽

2020 ◽

pp. 707-725

Author(s):

Sujata Dash

Keyword(s):

Feature Extraction ◽

Feature Selection ◽

Microarray Data ◽

Classification Model ◽

Rotation Forest ◽

Ensemble Technique ◽

Basic Characteristics ◽

Microarray Datasets ◽

Feature Selection Techniques

Efficient classification and feature extraction techniques pave an effective way for diagnosing cancers from microarray datasets. It has been observed that the conventional classification techniques have major limitations in discriminating the genes accurately. However, such kind of problems can be addressed by an ensemble technique to a great extent. In this paper, a hybrid RotBagg ensemble framework has been proposed to address the problem specified above. This technique is an integration of Rotation Forest and Bagging ensemble which in turn preserves the basic characteristics of ensemble architecture i.e., diversity and accuracy. Three different feature selection techniques are employed to select subsets of genes to improve the effectiveness and generalization of the RotBagg ensemble. The efficiency is validated through five microarray datasets and also compared with the results of base learners. The experimental results show that the correlation based FRFR with PCA-based RotBagg ensemble form a highly efficient classification model.

Download Full-text

Feature Selection Using Random Forest Algorithm to Diagnose Tuberculosis From Lung CT Images

AI Innovation in Medical Imaging Diagnostics - Advances in Medical Technologies and Clinical Practice ◽

10.4018/978-1-7998-3092-4.ch005 ◽

2021 ◽

pp. 92-100

Author(s):

Beaulah Jeyavathana Rajendran ◽

Kanimozhi K. V.

Keyword(s):

Feature Selection ◽

Random Forest ◽

The Body ◽

Support Vector ◽

Feature Descriptor ◽

Feature Sets ◽

Modified Particle Swarm Optimization ◽

Tuberculosis Disease ◽

Optimal Feature ◽

Lung Ct

Tuberculosis is one of the hazardous infectious diseases that can be categorized by the evolution of tubercles in the tissues. This disease mainly affects the lungs and also the other parts of the body. The disease can be easily diagnosed by the radiologists. The main objective of this chapter is to get best solution selected by means of modified particle swarm optimization is regarded as optimal feature descriptor. Five stages are being used to detect tuberculosis disease. They are pre-processing an image, segmenting the lungs and extracting the feature, feature selection and classification. These stages that are used in medical image processing to identify the tuberculosis. In the feature extraction, the GLCM approach is used to extract the features and from the extracted feature sets the optimal features are selected by random forest. Finally, support vector machine classifier method is used for image classification. The experimentation is done, and intermediate results are obtained. The proposed system accuracy results are better than the existing method in classification.

Download Full-text

COMPARISON OF TYPE-2 FUZZY CLUSTERING-BASED CASCADE CLASSIFIER MODELS FOR ECG ARRHYTHMIAS

Biomedical Engineering Applications Basis and Communications ◽

10.4015/s1016237214500756 ◽

2014 ◽

Vol 26 (06) ◽

pp. 1450075

Author(s):

Rahime Ceylan ◽

Yüksel Özbay ◽

Bekir Karlik

Keyword(s):

Feature Extraction ◽

Fuzzy Clustering ◽

Classification Algorithms ◽

Ecg Signal ◽

Classification Models ◽

Cascade Classifier ◽

Training Set ◽

Feature Sets

The aim of this study is to present a comparison of the novel cascade classifier models based on fuzzy clustering and feature extraction techniques according to efficiency. These models are composed of three subsystems: The first subsystem is constituted by fuzzy clustering technique to choose the best patterns that ideally show its class attributes in dataset. The second subsystem consists of discrete wavelet transform (DWT) which realizes feature extraction procedure on selected patterns by using fuzzy c-means clustering. The last subsystem implements the classification of extracted features for each pattern using classification algorithm. In this paper, type-2 fuzzy c-means (T2FCM) clustering is used in the first subsystem of the proposed classification models and the new training set is obtained. In the second subsystem, the features of the obtained new training set are extracted with DWT; hence, three different feature sets along with different number of features are formed using Daubechies-2 wavelet function. In the last subsystem of the model, the feature sets are classified using classification algorithm. Here, two different classification algorithms, neural network (NN) and support vector machine (SVM), are used for comparison. Thus, two classification models are implemented and named as T2FCWNN (classifier is NN) and T2FCWSVM (classifier is SVM), respectively. This proposed classifier models have been applied to classify electrocardiogram (ECG) signals. One of the goals of this study is to present a fast and efficient classifier. For this reason, high accuracy rate is been aimed for classification of RR intervals in ECG signal. So, we have utilized T2FCM and WTs to improve the performance of the classification algorithms. Both training and testing set for classifier models have included 12 ECG signal classes. Well-known back propagation algorithm has been used for training of neural networks (NNs). The best testing results have been obtained with 99% recognition accuracy with T2FCWNN-2.

Download Full-text