|
|
April 2022 Issue Vol.12 No.4
PREDICTION AND CLASSIFICATION USING RANDOM SUBSPACE CONDITIONAL PROBABILITIES TECHNIQUE FOR HEALTHCARE DATASETS
https://ia601500.us.archive.org/17/items/2022_ijitce/vol12no401.pdf
S. N. Santhalakshmi
Ph.D Research Scholar (Part-time), Department of Computer Science,
Nandha Arts and Science College, Erode, Tamil Nadu, India.
Dr. S. Prasath
Assistant Professor & Research Supervisor, School of Computer Science,
VET Institute of Arts and Science (Co-Edu) College, Erode, Tamil Nadu, India.
Abstract:
Today there is increase in society suffering from Diabetes disease and this number is rising continuously. Diabetes is a chronic disease that leads to numerous amount of death each year. Untreated diabetes troubles the proper functionality of other organs in mankind. Hence, identifying diabetes is very important to save the human life. Data mining is the process of analyzing data based on different factors and summarizing it into useful information. Prediction is one of the
mostly used techniques in medical data mining. The main aim of this work is to discover new patterns to provide meaningful and useful information for the public. The data are collected from clinic as well as in repository. The clinical data have some unknown values. Data mining techniques are applied to healthcare datasets to explore satisfactory methods and techniques in order to extract useful patterns with high accuracy with unknown values also. Generally, decision
tree classifies the data it won’t predict and this paper proposes an enhanced method which boosts up and develops the traditional classification algorithm for prediction. The proposed method is evaluated in WEKA tool with proper evaluation measures to confirm its efficiency.
Keywords: Classification, Prediction, Decision tree, Random subspace, Conditional Probabilities, Random forest, MLP.
