The Financial Data of Anomaly Detection Research Based on Time Series

Abstract

The rapid development of financial markets make the financial data variability and unpredictability, the abnormal fluctuations in financial data often contain important information. Financial data is generated over time, so the time series mining method widely used in the financial data of anomaly detection. The traditional time series of anomaly detection method is to find out one of the biggest point of outliers from a series of randomly generated numbers. It often do not consider the time sequence of time series, and the concern is not only a point of unusual data, but the abnormal sub-sequences in the time series anomaly detection. This paper presents a method that combines the activity and density of time series. It uses the time sequence of time series and sub-sequences' features effectively to discover the anomalies. Experimental results show that this method can be more effective and accurate to find the anomalies from the time series of financial data.