learning_analytics_tools_NPTEL_toooo.pptx

LAT

1. Ramkumar Rajendran IIT Bombay | [email protected] Learning Analytics Tools: Process Mining

2. Process Mining Process models from temporal data Different algorithm to develop process models Alphaminer Heuristic Miner Fuzzy Miner Van der Aalst, W., Weijters, T., & Maruster, L. (2004). Workflow mining: Discovering process models from event logs. IEEE Transactions on Knowledge and Data Engineering, 16(9), 1128–1142. Van

3. Process Mining Analyzes temporal sequence data to develop a process models that contain a set of nodes (events or actions) and edges (transition between actions/nodes) To develop an abstract process model Two key metrics, significance, correlation 2

4. Process Mining Significance is measured for both nodes and edges, by the relative importance of their occurrence compared to the total occurrence For example, the nodes or edges that occur more frequently are considered as more significant Correlation is also measured for both nodes and edges, by analyzing how two events are closely related For example, the two nodes that co-occur more frequently compared to other events is considered to be more correlated. 2

5. 2 Temporal Data is analysed to develop process model Sequence of actions: A, A, B, A, B, A, A, B Create state transition diagram Process Mining Activity

6. Fuzzy Miner Temporal Data is analysed to develop process model Sequence of actions: A, A, B, A, B, A, A, B Significance of action “A” = 1 and “B” = 0.6 Correlation: A  B = 3/5 = 0.6 A 1 B 0.6

7. Fuzzy Miner Process Model – Fuzzy Miner Highly significant Nodes are Preserved Less significant nodes that are highly correlated are aggregated into clusters Less significant nodes with low correlation with other nodes are dropped

8. ProM – Fuzzy Miner The abstraction level of PM modified by changing Node cutoff – to remove nodes whose significance is lower than cutoff Edge cutoff – to filter edges whose utility values is below the cutoff Utility ratio ( ur ) – 0 to 1

9. Rajendran, R., Munshi, A., Emara, M., & Biswas, G. (2018). A temporal model of learner behaviors in OELEs using process mining. In  Proceedings of ICCE  (pp. 276-285). https :// pdfs.semanticscholar.org/2ebf/7b20883f963f991ec44709e5052f8f854273.pdf Example 8

10. Process Mining ProM tool Summary 8

11. Thank You 8