Outlier Detection and Influential Point Observation in Linear Regression Using Clustering Techniques in Financial Time Series Data

The modern computing technology makes data gathering and storage easier. This creates new range of problems and challenges for data analysis. Detection of outliers in time series data has gained much attention in recent years. We present a new approach based on clustering techniques for outlier. The Expectation Maximization clusters (EM-Cluster) a l g o r i t h m is used to find the “optimal” parameters of the distributions that maximize the likelihood function. Regression based outlier technique is used to detect influence point. The analysis of outliers and influential points is an important step of the regression diagnostics. Several indicators are used for identifying and analyzing outliers. The proposed approach gave effective results within optimum time and space when applied to synthetic data set. This paper investigates the outliers, volatility clustering and risk-return trade-off in the Indian stock markets NSE Nifty and BSE SENSEX. Engle's ARCH Test and AR (1)-EGARCH (p, q)-in-Mean model were employed to examine the objective of the study. It revealed that volatility is persistent and there is leverage effect in the Indian stock markets.