Combining a gravitational search algorithm, particle swarm optimization, and fuzzy rules to improve the classification performance of a feed-forward neural network

Our plain-English summary of Combining a gravitational search algorithm, particle swarm optimization, and fuzzy rules to improve the classification performance of a feed-forward neural network, by M.L. Huang, Y.C. Chou (Computer Methods and Programs in Biomedicine, 2019).

The study

• Authors: M.L. Huang, Y.C. Chou
• Journal: Computer Methods and Programs in Biomedicine
• Published: 2019;180:105016
• Cited by: 3 articles, as counted by NIH iCite
• Read it: PubMed · Publisher

Abstract

Background And Objective: A feed-forward neural network (FNN) is a type of artificial neural network that has been widely used in medical diagnosis, data mining, stock market analysis, and other fields. Many studies have used FNN to develop medical decision-making systems to assist doctors in clinical diagnosis. The aim of the learning process in FNN is to find the best combination of connection weights and biases to achieve the minimum error. However, in many cases, FNNs converge to the local optimum but not the global optimum. Using open disease datasets, the purpose of this study was to optimize the connection weights and biases of the FNN to minimize the error and improve the accuracy of disease diagnosis.

Method: In this study, the chronic kidney disease (CKD) and mesothelioma (MES) disease datasets from the University of California Irvine (UCI) machine learning repository were used as research objects. This study applied the FNN to learn the features of each datum and used particle swarm optimization (PSO) and a gravitational search algorithm (GSA) to optimize the weights and biases of the FNN classifiers based on the algorithms inspired by the observation of natural phenomena. Moreover, fuzzy rules were used to optimize the parameters of the GSA to improve the performance of the algorithm in the classifier.

Results: When applied to the CKD dataset, the accuracies of PSO and GSA were 99%. By using fuzzy rules to optimize the GSA parameter, the accuracy of fuzzy-GSA was 99.25%. The accuracies of the combined algorithms PSO-GSA and fuzzy-PSO-GSA reached 100%. In the MES disease dataset, all methods exhibited good performance with 100% accuracy.

Conclusions: This study used PSO, GSA, fuzzy-GSA, PSO-GSA, and fuzzy-PSO-GSA on CKD and MES disease datasets to identify the disease, and the performance of different algorithms was explored. Compared with other methods in the literature, our proposed method achieved higher accuracy.

Copyright © 2019 Elsevier B.V. All rights reserved.

In plain English

This was a computer-science methods study using existing records. The authors tested ways of training an artificial neural network to sort records as sick or not. They used two public datasets from the University of California Irvine repository. One covered chronic kidney disease: 400 entries from India, 250 with the disease and 150 normal. The other, a mesothelioma dataset from Dicle University, Turkey, had 228 benign-tumor and 96 malignant-tumor entries. Its 34 variables included asbestos exposure and duration of asbestos exposure. The study did not examine asbestos as a cause of disease. The authors compared five training methods: particle swarm optimization (PSO) alone, a gravitational search algorithm (GSA) alone, GSA tuned with fuzzy rules, PSO and GSA combined, and that combination tuned with fuzzy rules. In the main comparison on the kidney dataset, accuracy ranged from 99% to 100%. With other settings, it fell as low as 98.50%. On the mesothelioma dataset, every method reached 100% accuracy under every setting tested. The authors concluded that their method was more accurate than others in the literature. They said experiments on other disease datasets are still needed.

We wrote this plain-English summary from the full article with the help of AI, and a second AI model checked each sentence against the article. It is not the authors' wording; the abstract and the article are their own account.

What the study did and found

• Design: Computational (machine learning) methods study comparing five ways of training a feed-forward neural network on existing public disease datasets
• Who was studied: Two public datasets. The chronic kidney disease dataset had 400 entries from India in 2015: 250 with the disease and 150 normal subjects. The mesothelioma dataset came from Dicle University School of Medicine in Turkey and had 228 entries for benign tumor and 96 entries for malignant tumor.
• Exposure: Asbestos was not studied as an exposure. The mesothelioma dataset included 'Asbestos exposure' (coded 0 or 1) and 'Duration of asbestos exposure' among its 34 input variables. No fiber types were given.
• Main results:
  • In the main comparison on the chronic kidney disease dataset, the PSO and GSA methods each had 99% accuracy.
  • In the main comparison on the chronic kidney disease dataset, the fuzzy-GSA method had 99.25% accuracy.
  • In the main comparison on the chronic kidney disease dataset, the combined PSO-GSA and fuzzy-PSO-GSA methods reached 100% accuracy.
  • On the chronic kidney disease dataset, the authors also varied the number of particles and the maximum number of iterations. The lowest accuracy reported in these tests was 98.50%, for the GSA method with 20 particles and 500 iterations.
  • On the mesothelioma dataset, all five methods had 100% accuracy.
  • On the mesothelioma dataset, accuracy stayed at 100% for all models when the number of particles and the maximum number of iterations were varied.
  • On the mesothelioma dataset, the authors' fuzzy-PSO-GSA method had the smallest mean square error of the methods compared (0.0053e-02).
  • The authors stated that their 100% results on the mesothelioma dataset were higher than those reported by three earlier studies of the same dataset.
• The authors' conclusion: The authors concluded that, compared with other methods in the literature, their proposed method achieved higher accuracy.
• Limitations the authors state:
  • The authors stated that experiments on other disease datasets are needed to confirm that the method is suitable.

Topics

Medical terms in this study, explained in our Lexicon: Mesothelioma.

Related studies

• Prediction of Radiation Pneumonitis With Machine Learning in Stage III Lung Cancer: A Pilot Study — M. Yakar, D. Etiz, et al. (2021)
• Identifying patients with suspected lung cancer in primary care: derivation and validation of an algorithm — J. Hippisley-Cox, C. Coupland (2011)
• Comparison of the ultra-low-dose Veo algorithm with the gold standard filtered back projection for detecting pulmonary asbestos-related conditions: a clinical observational study — M. Tekath, F. Dutheil, et al. (2014)
• Deep learning-based classification of mesothelioma improves prediction of patient outcome — P. Courtiol, C. Maussion, et al. (2019)
• A risk-based soft sensor for failure rate monitoring in water distribution network via adaptive neuro-fuzzy interference systems — M. Gheibi, R. Moezzi, et al. (2023)
• Convolutional Neural Networks for Segmentation of Malignant Pleural Mesothelioma: Analysis of Probability Map Thresholds (CALGB 30901, Alliance) — M. Shenouda, E. Gudmundsson, et al. (2023)

Browse return to medical research.