Page 43 - 2023-Vol19-Issue2

39 | Andreswari, Millenia, Rizky, Haniyah & Mufti

The process model with too many relations makes it difficult log conformance value with a combination of the Preserve
for the reader to analyze the graph because it is difficult to Threshold, Ratio Threshold, Edge Cutoff, Utility Ratio and
draw the main conclusions obtained. Spaghetti processes Node Cutoff parameters. Based on research conducted by
can be simplified through a graph simplification process by [10], the process model with the best conformance is carried
applying the Heuristic Miner and Fuzzy Miner algorithms at out by preserving threshold 0.05, utility ratio 0.85, edge cutoff
the conformance stage using ProM. 0.05, and node cutoff 0.05 [3].

B. Conformance Checking
Conformance checking is the validation stage in the initial

process model generated against the event log data used. The
process model generated in the discovery stage can be sim-
plified by the Fuzzy Miner and Heuristic Miner algorithms to
avoid spaghetti processes through the simplification process.
The application of the Heuristic Miner in ProM is carried
out using an interactive Data-aware Heuristic Miner (iDHM),
while the application of the Fuzzy Miner is carried out using
Mine for a Fuzzy Model (Fig. 3).

Fig. 4. Fuzzy Miner process model with scenario

Fig. 3. Fuzzy Miner process model without scenario

C. Mine for a Fuzzy Model Blue labels indicate clusters and yellow labels indicate
In the process modeling with the Fuzzy Miner algorithm, cluster elements or log activity performed (Fig. 4). The
resulting process model shows that there are 5 main clusters
a process model without scenario is made and several work- produced, namely cluster 42, cluster 47, cluster 48, cluster 49,
ing scenarios are applied to determine the best conformance and cluster 50. Activities that have a high correlation with
(Fig 4). Scenario implementation is done by finding the best other activities are assignments from cluster 42 (Table III).