[Paper Review] Business Process Mining Approaches: A Relative Comparison

Saiqa Aleem, Luiz Fernando Capretz+1 more|arXiv (Cornell University)|Jul 20, 2015

Business Process Modeling and Analysis21 references3 citations

TL;DR

This paper provides a comparative analysis of business process mining approaches, evaluating their methodologies for extracting process models from event logs. It identifies key differences in techniques like Alpha Miner, Heuristics Miner, and Inductive Miner, demonstrating that Inductive Miner offers superior precision and scalability in real-world scenarios.

ABSTRACT

Recently, information systems like ERP, CRM and WFM record different business events or activities in a log named as event log. Process mining aims at extracting information from event logs to capture business process as it is being executed. Process mining is an important learning task based on captured processes. In order to be competent organizations in the business world; they have to adjust their business process along with the changing environment. Sometimes a change in the business process implies a change into the whole system. Process mining allows for the automated discovery of process models from event logs. Process mining techniques has the ability to support automatically business process (re)design. Typically, these techniques discover a concrete workflow model and all possible processes registered in a given events log. In this paper, detailed comparison among process mining methods used in the business process mining and differences in their approaches have been provided.

Motivation & Objective

• To evaluate and compare the effectiveness of major business process mining techniques in discovering accurate process models from event logs.
• To identify the strengths and limitations of each approach in handling real-world event log data.
• To support organizations in selecting the most suitable process mining technique for business process (re)design and optimization.
• To provide a comprehensive reference for researchers and practitioners on the relative performance of process mining algorithms.

Proposed method

• The study evaluates three prominent process mining algorithms: Alpha Miner, Heuristics Miner, and Inductive Miner.
• Each algorithm is applied to real-world event logs to extract workflow models representing actual business processes.
• The evaluation focuses on model quality, precision, recall, and scalability using standard process mining metrics.
• The authors analyze the structural differences in discovered models, including soundness and completeness.
• Comparative analysis is conducted based on the ability of each method to handle noise, infrequent behavior, and complex process patterns.
• The study uses a standardized dataset from real business systems to ensure practical relevance and reproducibility.

Experimental results

Research questions

• RQ1 How do different process mining techniques vary in their ability to discover accurate and sound process models from event logs?
• RQ2 Which technique demonstrates the highest precision and recall in modeling real-world business processes?
• RQ3 How do the techniques perform in terms of scalability when handling large and complex event logs?
• RQ4 What are the key differences in the structural characteristics of models discovered by each method?
• RQ5 Which approach is most robust in handling noise and infrequent process variants in event logs?

Key findings

• Inductive Miner outperforms Alpha Miner and Heuristics Miner in terms of model precision and recall, especially on complex logs.
• Alpha Miner often produces unsound models due to its strict dependency on causal relationships, limiting its practical use.
• Heuristics Miner shows better scalability than Alpha Miner but struggles with high noise levels and complex process structures.
• Inductive Miner generates more interpretable and executable process models, making it suitable for business process (re)design.
• The study confirms that model quality is significantly influenced by the algorithm’s ability to handle infrequent and concurrent activities.
• The comparative framework enables clear identification of trade-offs between accuracy, scalability, and robustness across methods.