Pengembangan Metode Local Outlier Factor (LOF) Berbasis Pruned Partisi Dengan Memperhatikan Jarak Ketercapaian Deteksi pada Data Stream

Abstract

Outlier detection is studied and applied in many domains. Outliers arise due to different reasons such as fraudulent activities, structural defects, and mechanical issues. The detection of outliers is a challenging task that can reveal system faults, fraud, and save people’s lives. Outlier detection techniques are often domain-specific. The main challenge in outlier detection relates to modelling the normal behaviour in order to identify abnormalities. The choice of model is important, i.e., an unsuitable data model can lead to poor results. This requires a good understanding and interpretation of the data, the constraints, and requirements of the domain problem. Outlier detection is largely an unsupervised problem due to unavailability of labelled data and the fact that labelled data is expensive. During the extraction process, the outsource may damage their original data set and that will be defined as the intrusion. To avoid the intrusion and maintain the anomaly detection in a high densely populated environment is another difficult task. For that purpose, Grid Partitioning for Outlier Detection (GPOD) has been proposed for high density environment. This technique will detect the outlier using the grid partitioning approach and density based outlier detection scheme. Initially, all the data sets will be split in the grid format. Allocate the equal amount of data points to each grid. Compare the density of each grid to their neighbor grid in a zigzag manner. Based on the response, lesser density grid will be detected as outlier function as well as that grid will be eliminated. This proposed Grid Partitioning for Outlier Detection (GPOD) has reduced the complexity and increases the accuracy.