Arnab Dey

Arnab Dey, Mononito Goswami, Joo Heung Yoon et al. · cmu

A significant proportion of clinical physiologic monitoring alarms are false. This often leads to alarm fatigue in clinical personnel, inevitably compromising patient safety. To combat this issue, researchers have attempted to build Machine Learning (ML) models capable of accurately adjudicating Vital Sign (VS) alerts raised at the bedside of hemodynamically monitored patients as real or artifact. Previous studies have utilized supervised ML techniques that require substantial amounts of hand-labeled data. However, manually harvesting such data can be costly, time-consuming, and mundane, and is a key factor limiting the widespread adoption of ML in healthcare (HC). Instead, we explore the use of multiple, individually imperfect heuristics to automatically assign probabilistic labels to unlabeled training data using weak supervision. Our weakly supervised models perform competitively with traditional supervised techniques and require less involvement from domain experts, demonstrating their use as efficient and practical alternatives to supervised learning in HC applications of ML.

Arnab Dey, Andrew Heger, Darin England

In this article, we propose a systematic approach for fire station location planning. We develop machine learning models, based on Random Forest and Extreme Gradient Boosting, for demand prediction and utilize the models further to define a generalized index to measure quality of fire service in urban settings. Our model is built upon spatial data collected from multiple different sources. Efficacy of proper facility planning depends on choice of candidates where fire stations can be located along with existing stations, if any. Also, the travel time from these candidates to demand locations need to be taken care of to maintain fire safety standard. Here, we propose a travel time based clustering technique to identify suitable candidates. Finally, we develop an optimization problem to select best locations to install new fire stations. Our optimization problem is built upon maximum coverage problem, based on integer programming. We further develop a two-stage stochastic optimization model to characterize the confidence in our decision outcome. We present a detailed experimental study of our proposed approach in collaboration with city of Victoria Fire Department, MN, USA. Our demand prediction model achieves true positive rate of 80% and false positive rate of 20% approximately. We aid Victoria Fire Department to select a location for a new fire station using our approach. We present detailed results on improvement statistics by locating a new facility, as suggested by our methodology, in the city of Victoria.