The work done in this chapter demonstrates how the main components of the SPMaAF framework and sets of algorithms described earlier in Chapters 3 and 4, respectively, fit and rely on each other in achieving the semantic enhancement of the discovered process models. This is done by representing the models discovered through the standard process mining techniques as a set of annotated terms that links to or references the concepts defined within ontologies. It permits the process analysts to formally represent and analyse the several information in the underlying knowledge-bases in a more efficient and yet accurate manner. Henceforth, the conceptualisation method or tactics is allied to semantic lifting of the process models.
TopAnnotation Of Fuzzy Learning Model
Indeed, the first step towards achieving the semantic annotation of any given model should be aimed at making use of process description languages/assertions to link elements in the models with concepts that they represent in a well-defined ontology. Using the learning process model as a case study, we demonstrate that the main purpose of the semantic annotation method must be to seek the equivalence between the concepts of the process models (e.g fuzzy models) derived by applying the fuzzy miner algorithm on the learning process logs and the concepts of the defined (learning) process domain ontologies. Apparently, this is done by making use of the process descriptions languages and/or notations (semantic annotation) to represent the extracted models.
To this end, in order to perform the semantic annotation of the learning process models which this work uses to illustrate the method throughout this book, and the application of the semantic reasoning - the work applies the following process mining technique especially as a way to achieve the target objectives defined in Phase 1 of the SPMaAF framework or yet Algorithm 1 (see: Chapter 4) as follows:
First, we analyse the extracted events log for the Learning process (Okoye et al, 2016) using the fuzzy miner (Günther & Van der Aalst, 2007). The outcome or result of applying the fuzzy miner algorithm is as shown in the following Figure 1a and 1b. Fundamentally, the method involves the extraction and automated modelling of the process history data (see: Figures 2 to 6) by submitting the resulting event streams format to the process mining environment in Disco (Rozinat & Gunther, 2012) to help in discovery (mapping) of the fuzzy model represented in Figures 1a and 1b.
In turn, the method provides us with reliable and extendible results and/or insights about the readily available datasets in order to further create a model that describes the individual traces (or sequence workflows) learned from the different activities based on the proven framework of the Fuzzy Miner (Gu¨nther & Van der Aalst, 2006).