Fuzzy Bat based Cluster Center Selection Algorithm (FBCCSA) Improved K-Means Algorithm for Type 2 Diabetes Mellitus Prediction
P.V. Sankar Ganesh1,  P. Sri Priya2

1P. V. Sankar Ganesh, Research Scholar, VISTAS Pallavaram, Chennai (Tamil Nadu), India.
2P. Sri Priya, Professor, Department of Computer Applications, VISTAS Pallavaram, Chennai (Tamil Nadu), India.
Manuscript received on 19 January 2020 | Revised Manuscript received on 02 February 2020 | Manuscript Published on 05 February 2020 | PP: 128-133 | Volume-8 Issue-4S5 December 2019 | Retrieval Number: D10031284S519/2019©BEIESP | DOI: 10.35940/ijrte.D1003.1284S519
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: A present days lot of people are preconception by diabetic mellitus because of its constantly increasing happening day to day life .The all the diabetic patients recognize care about health quality or risk aspect before identify the diagnosis. Gestational diabetes are affected Women’s during pregarency period. Earlier stage identified to solve the problems. Modern days data mining techniques are introduced for type 2 diabetics melitius.This proposed techniques to solve the problem and have accuracy of the prediction model get used for to more dataset. The newly designed K-means algorithm is predicted resolve the problem. Fuzzy bat algorithm also known as Fuzzy Bat based Cluster Center Selection Algorithm. Initiallycenter values is produced randomly and directly affect the result of clusters. The proposed Fuzzy bat algorithm center values are selected by measuring the difference between the diabetes samples and the within cluster sum of squared errors. If the distance is smaller than the clusters are formed as well as center are also selected from diabetes samples. An Improved K-means Algorithm are to be combined to FBCCSA.This Proposed Paper improved k means algorithm with Logistic Regression (LR) algorithm for classification. A Sample Dataset were Collected to determine the proposed technique, which were taken from the University of California Irvine (UCI) for Analysis. LR model has provides higher accuracy of prediction than those of other methods such as Hybrid prediction model(HPM)and Logistic Regression (LR).The proposed model guarantees that the quality of dataset is adequate and results show good performance. As a result, the model is shown to be useful for the realistic health administration of diabetes.
Keywords: Data Mining, Diabetes Mellitus, Clustering, Fuzzy Bat Algorithm (FBA), Logistic Regression (LR), University of California Irvine (UCI), Pima Indians Diabetes Dataset,
Scope of the Article: Clustering