Refining the Noisy Candidates in Learning Data for Improving Classification Performance

The data mining is a technique are used to learn a pattern and identify the pattern among a huge amount of data
using computer based algorithms or programs. In this technique the two main technique of learning is found first
supervised and second the unsupervised. But the quality of learning is depends upon the amount of data and
quality of learning pattern. Therefore, in order to improve the quality of learning data the pre-processing is
performed on data. In this paper a review on the pre-processing techniques and the data quality enhancement
techniques is reported. Further the key issues and challenges are addressed for improving the learning data
quality. Finally for resolving the issues a new technique is proposed in this paper and their future extension of the
work is also provided