Data Classification Using KNN-fuzzy Method Optimized By A Genetic Algorithm
Free (open access)
J. L. A. Rosa & N. F. F. Ebecken
Data classification using KNN-fuzzy method optimized by a genetic algorithm J. L. A. Rosa & N. F. F. Ebecken COPPE / Universidade Federal do Rio de Janeiro /Brazil Abstract This paper presents a classification method based on the KNN-Fuzzy classification algorithm, supported by a Genetic Algorithm. It discusses how to consider data classification according to the Fuzzy logic and its consequences in Data Mining predictions. Analyses are made upon the results obtained in the classification of several databases in order to demonstrate the proposed approach. 1 Introduction Data classification is one of the most used data mining tasks. ANDERBERGS [l] show many examples of classification in different fields as religion, etnia, type of job, style of clothes, etc., and also defines classification as being the process or act to associate a new item or comment to a category. As example, a person can be classified, according some attributes: sex (female or male), nationality (country where it was born), naturalness (state where it was born), instruction degree (illiterate or not), height (low, high). Small amounts of attributes, allow a simple and direct classification by principal component analysis and graphical analysis. Most of the knowledge discovery techniques are based on mathematical statistics and Machine Learning fundamentals . Many classification methods are used and some of them are described in .