time series; data mining; knowledge discovery; trend extremum representation. Abstract. In recent years, there has been an explosion of interest in mining time series databases. In this paper, we make some attempts to mining process industrial time series. As with most computer science problems, representation of the data is the key to efficient and effective solutions. We introduce a novel algorithm, Trend Extremum Representation, which is empirically proved to be superior to Piecewise Linear Representation and Important Points Representation in manipulating large-scale industrial data. Then, subsequent mining procedure is undertaken. Through clustering analysis and association rule discovery, several useful rules are derived for differentiating normal and abnormal events in everyday operations.
A New Method for Piecewise Linear Representation of Time Series Data
