2018/03/28 by Djordje Gligorijevic, Gligorijevic, Djordje, Jelena Stojanovic +3
Computer Science · Health Professions · Medicine · #Artificial Intelligence in Healthcare #Data-Driven Disease Surveillance #FOS: Computer and information sciences #Machine Learning (cs.LG) #Machine Learning (stat.ML) #Machine Learning in Healthcare
paper · pdf · doi:10.48550/arxiv.1803.11462
openalex publication_date 2018/03/28 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
For highly sensitive real-world predictive analytic applications such as\nhealthcare and medicine, having good prediction accuracy alone is often not\nenough. These kinds of applications require a decision making process which\nuses uncertainty estimation as input whenever possible. Quality of uncertainty\nestimation is a subject of over or under confident prediction, which is often\nnot addressed in many models. In this paper we show several extensions to the\nGaussian Conditional Random Fields model, which aim to provide higher quality\nuncertainty estimation. These extensions are applied to the temporal disease\ngraph built from the State Inpatient Database (SID) of California, acquired\nfrom the HCUP. Our experiments demonstrate benefits of using graph information\nin modeling temporal disease properties as well as improvements in uncertainty\nestimation provided by given extensions of the Gaussian Conditional Random\nFields method.\n