Service Analytics
Permanent URI for this collection
Browse
Recent Submissions
Item Customizing Context: Discovering the Optimal Integration of Context Data to Elevate ML-Driven Automated Support Ticket Classification(2025-01-07) Fuchs, Simon; Wiehl, Nathalie; Wittges, Holger; Krcmar, HelmutContext-aware Machine Learning (ML) is a well-established subfield of ML research. Here, the idea is to consider the context of an ML problem to enhance model performance. Here, the decisive factor is precise implementation of representative context data. The field of ML-driven support ticket classification has yet only scratched the potential of context awareness. Traditional text-only classification often lack accuracy, posing a challenge to the deployment of ML-automated support ticket classifiers. To address this, we explore four approaches in this paper integrating structured contextual data with textual descriptions. We found that correct incorporation of context data significantly enhances accuracy. We evaluate the effect of text embedding on model performance, highlighting the need for thoughtful data integration strategies. We demonstrate the effectiveness of context-aware approaches and explores alternative text preprocessing techniques. Closing, we discuss our findings – especially, the impact of data handling, data imbalance and interpretability on the overall automation project.Item Utilizing Data Fingerprints for Privacy-Preserving Algorithm Selection in Time Series Classification: Performance and Uncertainty Estimation on Unseen Datasets(2025-01-07) Böcking, Lars; Müller, Leopold; Kühl, NiklasThe selection of algorithms is a crucial step in designing AI services for real-world time series classification use cases. Traditional methods such as neural architecture search, automated machine learning, combined algorithm selection, and hyperparameter optimizations are effective but require considerable computational resources and necessitate access to all data points to run their optimizations. In this work, we introduce a novel data fingerprint that describes any time series classification dataset in a privacy-preserving manner and provides insight into the algorithm selection problem without requiring training on the (unseen) dataset. By decomposing the multi-target regression problem, only our data fingerprints are used to estimate algorithm performance and uncertainty in a scalable and adaptable manner. Our approach is evaluated on the 112 University of California riverside benchmark datasets, demonstrating its effectiveness in predicting the performance of 35 state-of-the-art algorithms and providing valuable insights for effective algorithm selection in time series classification service systems, improving a naive baseline by 7.32% on average in estimating the mean performance and 15.81% in estimating the uncertainty.Item Introduction to the Minitrack on Service Analytics(2025-01-07) Setzer, Thomas; VöSsing, Michael; Kühl, Niklas; Fromm, Hansjoerg