作者: E. Chandra Blessie , E. Karthikeyan , V. Thavavel
DOI: 10.1007/978-3-642-28926-2_60
关键词:
摘要: Feature selection is one of the important issues in machine learning. Some RELIEF based algorithms are considered as most successful for assessing quality features. RELIEF-DISC which was shown to be efficient estimating features, cannot handle incomplete data continuous In this paper, we propose an extended algorithm by introducing a new approach deal with noisy and sets. We investigate performance Decision Tree classifier imputing missing values algorithm. The datasets taken from UCI ML Repository.