Improve the classifiers efficiency by handling missing values in diabetes dataset using WEKA filters

Abstract

The In the health care sector Data Mining tools and machine learning algorithms play an important role. Artificial intelligence (AI) is a superset of Machine Learning (ML). ML is an application of Artificial intelligence and makes the systems automatically learn themselves and improve from experience without being explicitly programmed. In this paper we have discussed the importance of data pre-processing techniques in machine learning predictive modeling. We performed the analysis of diabetes dataset with the help of WEKA tool. Weka tool is an open source machine learning tool. Data pre-processing techniques are used to handle the missing values, noisy and outliers in the dataset. In this paper we have applied filters either to remove missing values and to replace missing values with mean, median or mode. Classification techniques are applied to classify the data, compare the results and evaluate data using 10-fold cross validation [1].