LLM-based Interface for Data-Driven Waiting Time Analysis of Business Processes
Name
Maksym Avramenko
Abstract
Recent advancements in process mining, a field at the intersection of data science and process management, have unveiled significant potential in analyzing business processes, particularly in analyzing waiting times between different activities to identify potential bottlenecks and inefficiencies. This thesis addresses a common limitation in existing process mining tools: their fixed analytical interfaces, which restrict dynamic interaction with data and limit user-driven analysis of the process. Fixed interfaces often hinder the ability to generate customized insights and adjust analyses in response to evolving business needs. To address this, the thesis proposes the integration of a Large Language Model-based (LLM) conversational interface with process mining tool, aiming to foster an interactive engagement with the data, allowing for personalized responses and actionable insights on reducing waiting times. This initiative aligns with recent explorations of integrating LLMs with process mining to enhance user interaction and understanding. The primary contribution of this thesis is the design, implementation, and evaluation of the LLM-based interface within process mining tool, which is anticipated to allow interactive navigation with the data plus the insights on how to reduce waiting times within the process with redesign suggestions based on event logs.
Graduation Thesis language
English
Graduation Thesis type
Master - Software Engineering
Supervisor(s)
Fredrik Payman Milani, Katsiaryna Lashkevich
Defence year
2024