A severe drawback of the ARIMA approach is its inability to identify complex characteristics. This limitation occurs because of the goal of characterizing all time series observations, the necessity of time series stationarity, and the requirement of residual normality and independence.
Data Mining [6, 7] is the analysis of data with the goal of uncovering hidden patterns. Data Mining encompasses a set of methods that automate the scientific discovery process. Its uniqueness lies in the types of problems addressed those with large data sets and complex, hidden relationships.
The TSDM framework innovates data mining concepts for analyzing time series data. In particular, a set of methods that reveals hidden patterns in time series data and overcomes limitations of traditional time series analysis techniques have been developed. The TSDM framework focuses on predicting events, which are important occurrences. This allows the TSDM methods to predict nonstationary, nonperiodic, irregular time series, including chaotic deterministic time series. The TSDM methods are applicable to time series that appear stochastic, but occasionally (though not necessarily periodically) contain distinct, but possibly hidden, patterns that are characteristic of the desired events.
They have been successfully applied to characterizing and predicting complex, nonperiodic, irregular, and chaotic time series events from both engineering and financial domains [1-4].
Previous work in the area of time series predictability has been done by Kaboudan [8, 9]. He developed a method using Genetic Programming (GP) to evolve the fittest regression model that best replicates a time series, and applied this method to predict stock prices. He also developed a metric that quantifies the ability of his GP algorithm to predict a time series, i.e., a measure of time series predictability.
Kaboudan's method is similar to the ARIMA methods mentioned above in that it characterizes and predicts all time series observations.