COMPARATIVE ANALYSIS OF DATA CLASSIFICATION TECHNIQUES BY PROCESSING ON MISSING VALUES INSTANCES SUBSTITUTION ON DATA FROM WEB REPOSITORY FOR KNOWLEDGE DISCOVERY

Abstract

An emerging technology extensively used in every field – Data mining that deals with the innovation and analysis of usage pattern(s) and relationships within trends leading to discover knowledge. This is a way of process to analyze a data from different viewpoints and summarized it into meaningful information. Data can be stored in a range of forms of databases and storage area. This can be superior way to extracting valuable knowledge for decision making. The main objective of this study is to show the classifying data with a reasonable accuracy for improving the performance in data mining by applying missing values instances substitution on a data. . The learning converses data mining techniques to process a medical data set and classify the importance of heart diseases. The massive amounts of data related heart diseases to be gathered, unfortunately that are not “mined” to determine hidden information for valuable decision by healthcare practitioners. A heart disease covers the different diseases that have an effect on the heart. Some categories to be include related heart diseases - Cardiomyopathy and Cardiovascular. The objective will be focused on getting the higher level percentage of accuracy. The implementation of this result carried out with the open source software environment Orange Canvas with three Classification Techniques – NB, KNN and SVM.