Paper Title
Multivariate Anomaly Detection on Stock Market Volumes

Abstract
Stock Market Volume is the quantity of shares or contracts traded in a stock market or a security amid a given timeframe. Anomaly identification is a significant information digging procedure for recognizing irregular patterns in time series data. There are several factors that make anomaly detection of multivariate time series more complicated. The investor needs to know the anomalies on stock market for efficient trading.The present work focuses on the detection of anomalies in the stock market volumes using multivariate Gaussian distribution. This work performs a comparative study of the anomalies of the companies listed in both BSE SENSEX and NIFTY 50 from ten different sectors for a selected time frame.Based on Average Price and Average Stock Volume, it was observed that BSE showed more anomalies compared to NSE for the case under consideration. The detection of the anomalies show significant but rare events, which can lead to apprehension in the investors’ mind. Also it is possible that the investor may use the anomalies for his advantage by exploiting the anomalies to earn superior returns. Keywords - Anomalies, Stock market volumes, Gaussian distribution, Average Price.