Naïve Bayes Based Multiple Parallel Fuzzy Reasoning Method For Medical Diagnosis


Ramanathan, Thirumalaimuthu Thirumalaiappan and Hossen, Md. Jakir and Sayeed, Md. Shohel (2022) Naïve Bayes Based Multiple Parallel Fuzzy Reasoning Method For Medical Diagnosis. Journal of Engineering Science and Technology., 17 (1). 0472-0490. ISSN 1823-4690

[img] Text
Naive Bayes Based Multiple Parallel.pdf
Restricted to Repository staff only

Download (827kB)


There are millions of sample medical cases recorded in many digital medical datasets that can be used by the data mining techniques for predicting any particular disease. Improving the classification accuracy in medical diagnosis based on patterns extracted from the available medical datasets is a challenging research problem as the medical datasets contain many complex patterns. In artificial intelligence, hybrid intelligent systems can support the data mining process to improve the accuracy of classification for medical diagnosis. Hybrid intelligent system is an integrated design of different artificial intelligence techniques such as neuro-fuzzy, genetic-fuzzy, etc., that has been successful in many applications such as data mining, computer vision, speech synthesis, etc. This paper proposes a hybrid intelligent method of integrating Naïve Bayes classifier and parallel fuzzy systems for the classification of type 2 diabetes. The proposed method employs multiple hybrid fuzzy systems in a parallel structure for effective classification on the data. The proposed method showed better classification accuracy of 90.26% when tested using the Pima diabetes dataset.

Item Type: Article
Uncontrolled Keywords: Diabetes mellitus, Fuzzy logic, Medical data mining, Naïve Bayes
Subjects: Q Science > QA Mathematics > QA1-43 General
Divisions: Faculty of Engineering and Technology (FET)
Faculty of Information Science and Technology (FIST)
Depositing User: Ms Nurul Iqtiani Ahmad
Date Deposited: 03 Mar 2022 01:05
Last Modified: 03 Mar 2022 01:05


Downloads per month over past year

View ItemEdit (login required)