DefinePK

Abstract

Data summarization in unrealistic or uncertain data streams is a basic concept in relational data sources. For outstanding data summarization on uncertain data stream evaluation with jumps of data streams environments. Traditionally single attribute summarization approach was introduced to define related instances to construct Uncertain One Class Classifier to summarize class instances perfectively. This framework kernel density based method to generate possible score to obtain each attribute with feasible data maintenance; UOCC also provides support vector machine (SVM) representation to summarization concept based on user’s preferences and user’s requirement in stored data source. It was generated possible score based on data instances. It is failed to support data exploration based on data attributes (characteristics) to utilize data instances with cluster relational data sets. So, we propose to develop Multi Attribute Grouping Method (MAGM) to define data summarization and portioned attribute selection for data exploration in uncertain data streams. MAGM defines a matrix to construct unidentified records into cluster in uncertain reliable data streams with attribute partitioning and feature selection. Our experimental results show effective data summarization with uniform user’s data exploration with their search histories from uncertain data streams with respect to time and other feature factors.