LR-MS: a mining system for local temporal rules

In recent years, there has been increased interest in using data mining techniques to extract temporal rules from time series database. Local temporal rules, which only a subsequence exhibits, are actually very common in practice. Efficient discovery of the time durations in which temporal rules are valid, i.e. rule distributions, could benefit KDD of many real applications. To support the interactive and efficient discovery of rule distributions, a mining system (LR-MS) has been designed and implemented. In this paper, we present the mining process of the system, which include preprocessing of raw data, generating of the rule sets of interest, dividing strategies for different mining interest, and generating of the representation of this knowledge. We have analyzed the behavior of our mining system with both synthetic data and real data. The results correspond with the definition of our problem and reveal a kind of novel knowledge.