An Improved CART Decision Tree for Datasets with Irrelevant Feature

摘要: Data mining tasks results are usually improved by reducing the dimensionality of data. This improvement however is achieved harder in case that data size moderate or huge. Although numerous algorithms for accuracy have been proposed, all assume inducing a compact and highly generalized model difficult. In order to address above said issue, we introduce Randomized Gini Index (RGI), novel heuristic function reduction, particularly applicable large scale databases. Apart from removing irrelevant attributes, our algorithm capable minimizing level noise greater extend which very attractive feature problems. We extensively evaluate its performance through experiments on both artificial real world datasets. The outcome study shows suitability viability approach knowledge discovery