Contributed Poster Presentations: Business and Economic Statistics Section

Process data commonly comes in the form of event logs, which record units going through said processes. An event log is made up of a unit identifier, activities through which units may pass, the time and date at which units go through these activities, as well as other possible descriptive variables.
The sequence of events pertaining to a given unit is called its "journey". Processes are subject to multiple variations, of multiple magnitude, and of multiple types. When these variations stack, or when a big enough variation occurs, concept drift is in effect. The difficulty lies in defining when variations are sufficient to alert for concept drift, as well as in a proper definition of concept drift with regards to this type of sequential data.
We therefore propose a full methodology to establish control charts, specifically "u" charts and "EWMA" charts, to follow complex drifts in process data, particularly in activity sequences. Our method monitors concept drift in an easy to implement way, and is easily interpretable even by non-statisticians. 

We introduce a new approach for decoupling trends (drift) and changepoints (shifts) in time series. Our locally adaptive model-based approach for robustly decoupling combines Bayesian trend filtering and machine learning based regularization. An over-parameterized Bayesian dynamic linear model (DLM) is first applied to characterize drift. Then a weighted penalized likelihood estimator is paired with the estimated DLM posterior distribution to identify shifts. We show how Bayesian DLMs specified with so-called shrinkage priors can provide smooth estimates of underlying trends in the presence of complex noise components. However, their inability to shrink exactly to zero inhibits direct changepoint detection. In contrast, penalized likelihood methods are highly effective in locating changepoints. However, they require data with simple patterns in both signal and noise. The proposed decoupling approach combines the strengths of both, i.e. the flexibility of Bayesian DLMs with the hard thresholding property of penalized likelihood estimators, to provide changepoint analysis in complex, modern settings. Our framework is outlier robust & can identify a variety of complex changes.