Data Summarization based on Multiple Attributes in Unreliable Categorical Data
Deena Babu Mandru1, N. SwapnaSuhasini2, S. Pavan Kumar Reddy3
1Deena Babu Mandru, Assistant Professor, Department of CSE, Malla Reddy Institute of Technology, Maisammaguda, Secunderabad (Telangana), India.
2N. Swapna Suhasini, Assistant Professor, Department of CSE, Malla Reddy Institute of Technology, Maisammaguda, Secunderabad (Telangana), India.
3S. Pavan Kumar Reddy, Assistant Professor, Department of CSE, Malla Reddy Institute of Technology, Maisammaguda, Secunderabad (Telangana), India.
Manuscript received on 15 October 2019 | Revised Manuscript received on 24 October 2019 | Manuscript Published on 02 November 2019 | PP: 2430-2434 | Volume-8 Issue-2S11 September 2019 | Retrieval Number: B12820982S1119/2019©BEIESP | DOI: 10.35940/ijrte.B1282.0982S1119
Open Access | Editorial and Publishing Policies | Cite | Mendeley | Indexing and Abstracting
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC-BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: Data summarization in preposterous or doubtful front page new streams is an integral production in relational story sources. For prosperous announcement summarization on between rock and hard place story cat and dog weather evaluation by all of the jumps of story streams environments. Traditionally one-class learning work summarization act was approved to translate the indistinguishable illustration and then constitute Undefined One Class Classifier (UOCC) by utilizing such class summarization effectively. This framework substance density based rule of thumb to inspire possible did a bang-up job to garner each chides mutually pragmatic front page new maintenance; UOCC furthermore provides support vector (SV) cross-section to summarization theory centered on user’s likings and article in the stored data source. It was produced potential database on data illustrations. It is unsuccessful to sponsor data distribution based on data characteristics to use data illustrations with cluster-based data sets. We proposed and implemented Enhanced Categorical Cluster Ensemble Approach (ECCEA) to handle data relations between different attributes to explore data from uncertain data. This approach consists of matrix to describe anonymous records into groups in indeterminate dependable data streams with attribute splitting and feature selection. Investigational outcomes of proposed approach give better and efficient cluster ensemble results with multi attributes in real time data sets.
Keywords: K-Means, Uncertain One Class Classifier, Cluster Ensemble Approach, Support Vector mechanism, Feature Representation.
Scope of the Article: Data Mining and Warehousing