Companies more and more rely on predictive services which are constantly monitoring and analyzing the available data streams for better service offerings. However, sudden or incremental changes in those streams are a challenge for the validity and proper functionality of the predictive service over time. We develop a framework which allows to characterize and differentiate predictive services with regard to their ongoing validity. Furthermore, this work proposes a research agenda of worthwhile research topics to improve the long-term validity of predictive services. In our work, we especially focus on different scenarios of true label availability for predictive services as well as the integration of expert knowledge. With these insights at hand, we lay an important foundation for future research in the field of valid predictive services.

Deployed machine learning models are confronted with the problem of changing data over time, a phenomenon also called concept drift. While existing approaches of concept drift detection already show convincing results, they require true labels as a prerequisite for successful drift detection. Especially in many real-world application scenarios-like the ones covered in this work-true labels are scarce, and their acquisition is expensive. Therefore, we introduce a new algorithm for drift detection, Uncertainty Drift Detection (UDD), which is able to detect drifts without access to true labels. Our approach is based on the uncertainty estimates provided by a deep neural network in combination with Monte Carlo Dropout. Structural changes over time are detected by applying the ADWIN technique on the uncertainty estimates, and detected drifts trigger a retraining of the prediction model. In contrast to input data-based drift detection, our approach considers the effects of the current input data on the properties of the prediction model rather than detecting change on the input data only (which can lead to unnecessary retrainings). We show that UDD outperforms other state-of-the-art strategies on two synthetic as well as ten real-world data sets for both regression and classification tasks.

Predictive services nowadays play an important role across all business sectors. However, deployed machine learning models are challenged by changing data streams over time which is described as concept drift. Prediction quality of models can be largely influenced by this phenomenon. Therefore, concept drift is usually handled by retraining of the model. However, current research lacks a recommendation which data should be selected for the retraining of the machine learning model. Therefore, we systematically analyze different data selection strategies in this work. Subsequently, we instantiate our findings on a use case in process mining which is strongly affected by concept drift. We can show that we can improve accuracy from 0.5400 to 0.7010 with concept drift handling. Furthermore, we depict the effects of the different data selection strategies.

This is a PDF file of an unedited manuscript that has been accepted for publication in the Communications of the Association for Information Systems. We are providing this early version of the manuscript to allow for expedited dissemination to interested readers. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered, which could affect the content. All legal disclaimers that apply to the Communications of the Association for Information Systems pertain.

No abstract