Learning from data to predict future symptoms of oncology patients

• DOI: 10.1371/journal.pone.0208808
• Publication Date: 2018-12-31T18:44:19Z
• Authors: Nikolaos Papachri...; Daniel Puschmann; Payam Barnaghi; Bruce Cooper; Xiao Hu; Roma Maguire; Kathi Apostolidis; Yvette P. Conley; Marilyn Hammer; Stylianos Katsara...; Kord M. Kober; Jon D. Levine; Lisa McCann; Elisabeth Patiraki; Eileen P. Furlong; Patricia A. Fox; Steven M. Paul; Emma Ream; Fay Wright; Christine Miaskowski
• Keywords: Depression; Regimen; Personalized Medicine

Effective symptom management is a critical component of cancer treatment. Computational tools that predict the course and severity these symptoms have potential to assist oncology clinicians personalize patient's treatment regimen more efficiently provide aggressive timely interventions. Three common inter-related in patients are depression, anxiety, sleep disturbance. In this paper, we elaborate on efficiency Support Vector Regression (SVR) Non-linear Canonical Correlation Analysis by Neural Networks (n-CCA) aforementioned between two different time points during cycle chemotherapy (CTX). Our results demonstrate methods produced equivalent for all three symptoms. These types predictive models can be used identify high risk patients, educate about their experience, improve timing pre-emptive personalized