Application of Retrieval of Similar Time Series with Linguistic Expressions to Diagnosis

Abstract

We have many kinds of data of time series. We understand them via linguistic expressions. We have proposed a method to express their global trend and local features in a natural language and defined a similarity degree of linguistic expressions for retrieving similar time series. In this paper, we apply the retrieval of similar timeseries to an alarm system of Waste Incineration plants, which alarms operators if the current time series are similar to stored time series of bad combustion air. We diagnose actual data from the plant.