Temporal mining for interactive workflow data analysis

摘要: In the past few years there has been an increasing interest in analysis of process logs. Several proposed techniques, such as workflow mining, are aimed at automatically deriving underlying models. However, current approaches only pay little attention on important piece information contained logs: timestamps, which used to define a sequential ordering performed tasks. this work we try overcome these limitations by explicitly including time extracted knowledge, thus making temporal first-class citizen process. This makes it possible discern between apparently identical executions that with different transition times consecutive tasks.This paper proposes framework for user-interactive exploration condensed representation groups given The is based use existing mining paradigm: Temporally-Annotated Sequences (TAS). These extracting patterns where each two events annotated typical emerges from input data. With TAS, represent sets frequent their times, factorizing operators built. condense according parallel or mutual exclusive executions. Lastly, rendered user via graph, namely Graph (TAG).The user, domain expert, allowed explore and alternative factorizations corresponding interpretations actual According choices, system discards retains certain hypotheses shows consequent scenarios resulting coresponding re-aggregation