Applicability of Process Mining Techniques in Business Environments

摘要: This thesis analyses problems related to the applicability, in business environments, of Process Mining tools and techniques. The first contribution is a presentation state art characterization companies, terms their "process awareness". The work continues identifying circumstance where can emerge: data preparation; actual mining; results interpretation. Other are configuration parameters by not-expert users computational complexity. We concentrate on two possible scenarios: "batch" "on-line" Mining. Concerning batch Mining, we investigated preparation problem proposed solution for identification "case-ids" whenever this field not explicitly indicated. After that, concentrated at mining time propose generalization well-known control-flow discovery algorithm order exploit non instantaneous events. usage interval-based recording leads an important improvement performance. Later on, report our users. We present approaches select "best" configuration: one completely autonomous; other requires human interaction navigate hierarchy candidate models. Concerning interpretation evaluation, metrics: model-to-model model-to-log. Finally, automatic approach extension model with social information, simplify analysis these perspectives. The second part deals algorithms on-line settings. formal definition problem, baseline approaches. two: adaptation, frequency counting algorithm; constitutes framework models which be used different kinds streams (stationary versus evolving).