Publication Date

Fall 2020

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science

First Advisor

Leonard Wesley

Second Advisor

Julia Newton

Third Advisor

William Andeopoulos


pancreatic cancer, evidential reasoning, biomarkers


In this project, an evidential reasoning model is built to amalgamate factors that could be used in early detection of pancreatic cancer. Our machine learning model outputs a probability of a given patient having prostate cancer based on various input variables. These variables include health history factors, such as smoking and medical history, technical artifacts, such as biopsy sequencing technology, and genomic biomarkers such as mutational, transcriptional and methylomic profiles, cfDNA, and copy number variation. The dataset used in this project is a part of The Cancer Genome Atlas (TCGA) project and was collected from the National Cancer Institute (NIH) Genomic Data Commons (GDC). The model is tested by varying input propositions and probability mass functions of input frames to create different combinations of input factors. Baseline prediction results in (0.084, 0.19) of not having pancreatic cancer. Prediction results were compared to the baseline prediction and a set of positive control expectations. For example, medium to high smoking history, medium to high drinking history with some cancer history will increase the posterior belief of a patient having pancreatic cancer to (0.091, 0.208). Presence of prognostic biomarkers will also increase the support for having pancreatic cancer, having medium impact DNA methylation and medium impact mRNA expression can lead the belief of having pancreatic cancer increase to (0.167, 0.273).