Interactive Context Search System for Exchange Rate Data Using Granularity-Dependent Similarity Calculation of Temporal Data

This paper proposes an interactive context search system for exchange rate data, which supports users to grasp the relationship between exchange rate and events. As the fluctuation of exchange rate is influenced by various kinds of events, it is important to grasp how a certain type of events will influence the fluctuation of exchange rate, and the type of events common to time periods having similar trends. Regarding the fluctuation of exchange rate / events as the context of the other, this paper defines two types of context search functions. One is summarizing events occurred in time period having similar fluctuation. Another is searching events with specified type along with corresponding time periods. In order to find time periods having similar fluctuation, this paper employs several similarity calculations for temporal data, among which suitable one is selected according to the granularity of fluctuation a user is focusing on. This paper shows the experimental results on similarity calculation with different granularities, as well as the results of experiment with test participants conducting exploratory analysis of fluctuation of exchange rate.