Analysing Performance Monitoring of Different Classification Techniques for Diebetic Data
No Thumbnail Available
Date
2016
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Data mining is the current research area to solve various problems and classification is one o f the main problem in
the field o f data mining. It allows users to analyze data from different dimensions or angles, categorize it, and
summarize the relationships identified. Classification is used to classify each item in a set o f data into one o f
predefined set o f classes or groups. In this paper we present algorithmic discussion ofJ48, Random tree, J48 Graft,
LAD and REP. Here the performance is compared for computing time, correctly classified instances, kappa
statistics, RMSE, MAE, RRSE, RAE and to find the error rate measurement for different classifiers using weka tool.
We have taken the classification o f data for diabetic patients data set is developed by collecting data from hospital
repository which has 1865 instances with different attributes. The dataset instances consist o f two categories o f
blood tests and urine tests. Weka tool classifies the data and is evaluated using 10 fold cross validation and the
results are then measured. It discovered that J48 performs better in most o f the cases.