Selecting an Efficient Classifier For The Given Dataset

Abstract

Over the past two decades, Machine Learning applications has spread to almost every fields and also the methods used for machine learning has converged from many sources. Current machine learning tools contain ex­cessive set of these methods but are lacking in providing assistance to select a suitable classifier because of the variations in the development and the disciplines. The main goal of meta learning is to provide this assistance to the end user who may not be expert in the field of machine learning but has to deal with the algorithms for the classification process. Here, in this research, the idea to design an assistant model using the meta learning techniques has been discussed. The survey of various meta learning tech­niques is done and have been explained in brief. Also, a hierarchical training approach using neural network, to get better results. The proposed work derives a metadataset and predicts the suitable classification algorithm for the new dataset.