Абстрактный

Improving Cluster Formulation to Reduce Outliers in Data Mining

Nancy Lekhi, Manish Mahajan

Existing studies in data mining focus on Outlier detection on data with single clustering algorithm mostly. There are lots of methods available in data mining to detect the outlier by making the clusters of data and then detect the outlier from them .Where outlier is the data item whose value falls outside the bounds in the sample data may indicate anomalous data. Outlier can be reduced if we improve the clustering .In this paper we proposed a hybrid algorithm that work not only on numeric data but also on text data. Our focus is to improve the cluster making so that the number of outliers can be reduce for that we can combine the clustering and classification techniques of data mining i.e. weighted k-mean and neural networks.