Search

Pig iron, the source for a variety of iron-based products, is traded in commodity markets. Therefore, enhanced productivity has significant economic implications for the producers. Pig iron is mainly produced inside of tall, vertical, thermodynamic reactors called blast furnaces that run 24 hours a day. The blast furnaces are too complex to model explicitly and are generally regarded as black boxes. In this study, we design, develop, and deploy novel machine learning models on a rich data sample covering more than 20 production variables spanning nine years of actual operational period, collected at one of the largest pig iron production plants in Brazil. We show that, given the blast furnace parameters, machine learning models are capable of unveiling novel insights by illuminating the black box and successfully predicting production levels at different configurations. These prediction models can be used as decision aids to improve production efficiencies. We also perform a sensitivity analysis of the trained models to identify and rank the input variables according to their relative importance. We present our findings, which are largely in line with the existing literature, and confirm their validity, practicality, and usefulness through consultations with subject matter experts. [ABSTRACT FROM AUTHOR]

Epilepsy is one of the most common brain disorders that greatly affects patients' quality of life and poses serious risks to their health. While the majority of the patients positively respond to the existing anti-epilepsy drugs, others who developed the refractory type of epilepsy show resistance against drug therapy and need to undergo advance treatments such as surgery. Given that identifying such patients is not a straightforward process and requires long courses of trial and error with anti-epilepsy drugs, this study aims at predicting those at-risk patients using clinical and demographic data obtained from electronic medical records. Specifically, the study employs several predictive analytics machine-learning methods, equipped with a novel approach for data balancing, to identify drug-resistant patients using their comorbidities and demographic information along with the initial epilepsy-related diagnosis made by their physician. The promising results we obtained highlight the potential use of machine-learning techniques in facilitating medical decisions and suggest the possibility of extending the proposed approach for developing a clinical decision support system for medical professionals. [ABSTRACT FROM AUTHOR]