Improved adaptive neuro-fuzzy inference system based on modified glowworm swarm and differential evolution optimization
- PDF / 940,150 Bytes
- 12 Pages / 595.276 x 790.866 pts Page_size
- 16 Downloads / 182 Views
(0123456789().,-volV)(0123456789(). ,- volV)
ORIGINAL ARTICLE
Improved adaptive neuro-fuzzy inference system based on modified glowworm swarm and differential evolution optimization algorithm for medical diagnosis Kishore Balasubramanian1
•
N. P. Ananthamoorthy2
Received: 23 April 2020 / Accepted: 4 November 2020 Springer-Verlag London Ltd., part of Springer Nature 2020
Abstract Medical diagnosis has seen a tremendous advancement in the recent years due to the advent of modern and hybrid techniques that aid in screening and management of the disease. This paper figures a predictive model for detecting neurodegenerative diseases like glaucoma, Parkinson’s disease and carcinogenic diseases like breast cancer. The proposed approach focuses on enhancing the efficiency of adaptive neuro-fuzzy inference system (ANFIS) using a modified glowworm swarm optimization algorithm (M-GSO). This algorithm is a global optimization wrapper approach that simulates the collective behavior of glowworms in nature during food search. However, it still suffers from being trapped in local minima. Hence in order to improve glowworm swarm optimization algorithm, differential evolution (DE) algorithm is utilized to enhance the behavior of glowworms. The proposed (DE–GSO–ANFIS) approach estimates suitable prediction parameters of ANFIS by employing DE–GSO algorithm. The outcomes of the proposed model are compared with traditional ANFIS model, genetic algorithm-ANFIS (GA-ANFIS), particle swarm optimization-ANFIS (PSO-ANFIS), lion optimization algorithm-ANFIS (LOA-ANFIS), differential evolution-ANFIS (DE-ANFIS) and glowworm swarm optimization (GSO). Experimental results depict better performance and superiority of the DE–GSO–ANFIS over the similar methods in predicting medical disorders. Keywords Adaptive neuro-fuzzy inference system Differential evolution Glowworm swarm optimization Neuro-ophthalmic disorders
1 Introduction Clinical datasets are widely used in predicting and managing many diseases like glaucoma, diabetic retinopathy, Parkinson’s disease, breast cancer, etc. Healthcare applications find extensive use of data mining approaches in analyzing the trends in subject’s records leading to overall improvement in healthcare. Prediction from the data mining process leads to systematic support in enhancing decision making [1]. The clinical data of the
& Kishore Balasubramanian [email protected] 1
Dr. Mahalingam College of Engineering and Technology, Pollachi, India
2
Hindusthan College of Engineering and Technology, Coimbatore, India
patients possess uncertainty in many ways, and hence, complete decision making with certainty is practically challenging. Certain diseases need early diagnosis and continual treatment as in the case of glaucoma for which computer-aided diagnosis (CAD) systems will be of much use to the clinicians in taking a second opinion before making a concrete decision and plan for treatment. In [2], a framework to get knowledge mining from patient’s clinical datasets has been presented. Numerous lit
Data Loading...
