Etude comparative des algorithmes dédiés à la classification

Abstract

Classification is the most popular data mining technique; it is used to categorize or classify information from a large data set in order to make predictions. There are many algorithms used for classification. But most of the existing algorithms are unstable in terms of accuracy and based on data content. This project aims to present a comparative study between classification algorithms, and to propose an ensemble approach in order to try to solve the problem of precision. The experiments and evaluations were carried out on medical databases and using the WEKA library. The results obtained are very satisfactory compared to the classification algorithms