• Ms.Ashwini Jadhav
• K.V.Metre

Local and global outliers detection are two important aspects in distance based outlier detection process. To analyze outliers from high speed data streams is challenging task. In existing outlier detection techniques, along with the incoming data stream points, previous data points are also stored for processing with the assumption of unbounded memory. To overcome this problem, memory efficient incremental outlier detection over data stream-(MiLOF) is also proposed in existing work. This technique preserves the summary of previous data points and finds local outliers. In this research work Memory efficient local outlier detection technique is proposed over data stream with feature extraction process. This technique preserves the summary of previous data points in each iteration and finds local outliers. This algorithm is divided is 3 phases namely: Summarization, Merging and Revised Insertion. To reach the memory constraint criteria summarization is used. To generate revised clusters merging phase is used. Whereas in revised insertion, LOF is identified and points are placed in cluster or marked as an outlier. Feature extraction process helps to reduce dimension in high dimensional dataset and hence to reduce memory for processing.

outlier detection, local outlier, clustering, density based outlier, data stream, feature extraction.

"Efficient Clustering Based Local Outlier Detection Over High Dimensional Streaming Data ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.5, Issue 8, page no.461-467, August-2018, Available :http://www.jetir.org/papers/JETIR1808066.pdf

"Efficient Clustering Based Local Outlier Detection Over High Dimensional Streaming Data ", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.5, Issue 8, page no. pp461-467, August-2018, Available at : http://www.jetir.org/papers/JETIR1808066.pdf