Wyniki wyszukiwania - BazTech

1

A comparative study on performance of basic and ensemble classifiers with various datasets

Gunakala Archana, Shahid Afzal Hussain

Applied Computer Science

|

2023

|

Vol. 19, no 1

107--132

EN

Classification plays a critical role in machine learning (ML) systems for processing images, text and high -dimensional data. Predicting class labels from training data is the primary goal of classification. An optimal model for a particular classification problem is chosen based on the model's performance and execution time. This paper compares and analyzes the performance of basic as well as ensemble classifiers utilizing 10-fold cross validation and also discusses their essential concepts, advantages, and disadvantages. In this study five basic classifiers namely Naïve Bayes (NB), Multi-layer Perceptron (MLP), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF) and the ensemble of all the five classifiers along with few more combinations are compared with five University of California Irvine (UCI) ML Repository datasets and a Diabetes Health Indicators dataset from Kaggle repository. To analyze and compare the performance of classifiers, evaluation metrics like Accuracy, Recall, Precision, Area Under Curve (AUC) and F-Score are used. Experimental results showed that SVM performs best on two out of the six datasets (Diabetes Health Indicators and waveform), RF performs best for Arrhythmia, Sonar, Tic-tac-toe datasets, and the best ensemble combination is found to be DT+SVM+RF on Ionosphere dataset having respective accuracies 72.58%, 90.38%, 81.63%, 73.59%, 94.78% and 94.01%. The proposed ensemble combinations outperformed the conven¬tional models for few datasets.

2

A Novel Approach for Coronary Artery Disease Diagnosis using Hybrid Particle Swarm Optimization based Emotional Neural Network

Shahid Afzal Hussain, Singh M. P.

Biocybernetics and Biomedical Engineering

|

2020

|

Vol. 40, no. 4

1568--1585

EN

Coronary artery disease (CAD) can cause serious conditions such as severe heart attack, heart failure, and angina in patients with cardiovascular problems. These conditions may be prevented by knowing the important symptoms and diagnosing the disease in the early stage. For diagnosing CAD, clinicians often use angiography, however, it is an invasive procedure that incurs high costs and causes severe side effects. Therefore, the other alternatives such as data mining and machine learning techniques have been applied extensively. Accordingly, the paper proposes a recent development of a highly accurate machine learning model emotional neural networks (EmNNs) which is hybridized with conventional particle swarm optimization (PSO) technique for the diagnosis of CAD. To enhance the performance of the proposed model, the paper employs four different feature selection methods, namely Fisher, Relief-F, Minimum Redundancy Maximum Relevance, and Weight by SVM, on Z-Alizadeh sani dataset. The EmNNs, with addition to the conventional weights and biases, uses emotional parameters to enhance the learning ability of the network. Further, the efficiency of the proposed model is compared with the PSO based adaptive neuro-fuzzy inference system (PSO-ANFIS). The proposed model is found better than the PSO-ANFIS model. The obtained highest average values of accuracy, precision, sensitivity, specificity, and F1-score over all the 10-fold cross-validation are 88.34%, 92.37%, 91.85%, 78.98%, and 92.12% respectively which is competitive to the known approaches in the literature. The F1-score obtained by the proposed model over Z-Alizadeh sani dataset is second best among the existing works.

3

Computational intelligence techniques for medical diagnosis and prognosis: Problems and current developments

Shahid Afzal Hussain, Singh M. P.

Biocybernetics and Biomedical Engineering

|

2019

|

Vol. 39, no. 3

638--672

EN

Diagnosis, being the first step in medical practice, is very crucial for clinical decision making. This paper investigates state-of-the-art computational intelligence (CI) techniques applied in the field of medical diagnosis and prognosis. The paper presents the performance of these techniques in diagnosing different diseases along with the detailed description of the data used. This paper includes basic as well as hybrid CI techniques that have been used in recent years so as to know the current trends in medical diagnosis domain. The paper presents the merits and demerits of different techniques in general as well as application specific context. This paper discusses some critical issues related to the medical diagnosis and prognosis such as uncertainties in the medical domain, problems in the medical data especially dealing with time-stamped (temporal) data, and knowledge acquisition. Moreover, this paper also discusses the features of good CI techniques in medical diagnosis. Overall, this review provides new insight for future research requirements in the medical diagnosis domain.