A comparison of classification methods as diagnostic system: A case study on skin lesions. (December 2016)

• Record Type: Journal Article
• Authors: Odeh, Suhail M.; Baareh, Abdel Karim Mohamed
• Is Part Of: Computer methods and programs in biomedicine. Volume 137(2016)
• Page Start: 311
• Page End: 319
• Publication Date: 2016-12
• Subjects: K nearest neighbour -- Genetic algorithm -- Artificial Neural Networks -- Adaptive Neuro-Fuzzy Inference System (ANFIS) -- Skin cancer -- Diagnosis system; Medicine -- Computer programs -- Periodicals; Biology -- Computer programs -- Periodicals; Computers -- Periodicals; Medicine -- Periodicals; Médecine -- Logiciels -- Périodiques; Biologie -- Logiciels -- Périodiques; Biology -- Computer programs; Medicine -- Computer programs; Periodicals; Electronic journals; 610.28
• DOI: 10.1016/j.cmpb.2016.09.012
• Languages: English
• ISSNs: 0169-2607

Abstract

Abstract: Background and objective: Numerous classification methods are currently available, but most of them were performed on different datasets. In this paper, different classification techniques were used for a diagnostic system on different skin lesions for the same data, which gives consistency for the data to have more accurate and better results. Methods: Four classification methods were proposed, a classical method based on K-Nearest Neighbor with Sequential Scanning selection technique for feature selection, a classical method with complex technique KNN with Genetic Algorithm, a complex method based on Artificial Neural Networks with Genetic Algorithm and an Adaptive Neuro-Fuzzy Inference System. Results: From the results obtained we can say that the performance of KNN with optimization of genetic algorithm for the feature selection was the best with an accuracy rate of 94%. The Adaptive Neuro-Fuzzy Inference System result was also good with an accuracy rate of 92%, and the other techniques' results were also satisfactory. Conclusion: The improvement on the performance of the classifier depends on the feature selection methods. In addition, the diagnosis system as a decision support tool could be used to increase the performance of human experts to make a correct decision.