Implementation of MD Heuristic Method for Classifying Numerical Data in Data Preprocessing

Abstract

Implementation of Data Mining with various methods will generate new knowledge and information in the form of rule which is very useful, but in implementation of data mining experience some weakness in data processing process. One weakness is that not all variable data can be easily processed, if the data is processed has a numerical value or data in the form of numbers then the division of classification is very difficult to do. If the classification of data is so many variables then the branch on the decision tree is also very much and result in less accurate results of rules or knowledge generated. MD Heuristic method can be used to classify numerical data or numbers that have a large range of data. By implementing the MD Heuristic Method in classifying the data in numerical form it can easily classify the data to make the data classification more precise and accurate, by calculating and averaging between the upper and lower bounds of the large data set. The use of MD Heuristic can facilitate preprocessing data with decision tree algorithm C4.5. The result of this research is detainee data at Labuhan Deli Detention Center, that is the data of age of detainee can be classified with more accurate to facilitate in process of data and more shorten rule or knowledge in result of decision tree image.