A Novel Cluster Based Outlier Elimination Algorithm

Outlier detection ensures that the data which is used to draw conclusions is consistent and reliable. The use of this technique has far reaching impact in a wide variety of different fields. In this paper, a new method for outlier detection is explored and presented. We will aim to show how outliers are detected using a sliding window of three points in the use case of High Frequency and Low Frequency time series data. This technique will be applied to this use case and an explanation of how the outliers were categorized will be provided. Experimental results will show that the algorithm created works successfully.