ABSTRACT

Outlier are those event which is drift too much after actions. So discovering outlier from a group of outline is a widespread problem in the area of data mining. The recognition of outlier can cause to locate some useful and meaningful knowledge. Previously outlier consider as noisy data, has now become frightful difficulty which has been uncovered in various domains of research. In this paper mainly focused on different kind of outlier and their detection approaches. This mainly contain classification of outlier and techniques which are classic outlier discovery method and spatial outlier discovery method. The classic outlier discovery method discover outlier in real transaction dataset, which is divided into statistical approach, distance approach, deviation approach and density approach. The spatial outlier method based on spatial datasets are separate from operation data, which are considered into spaced and graph approach. Finally, the application of outlier discovery approach.