Predicting Pancreatic Cancer Using Support Vector Machine

Abstract

This report presents an approach to predict pancreatic cancer using Support Vector Machine Classification algorithm. The research objective of this project it to predict pancreatic cancer on just genomic, just clinical and combination of genomic and clinical data. We have used real genomic data having 22,763 samples and 154 features per sample. We have also created Synthetic Clinical data having 400 samples and 7 features per sample in order to predict accuracy of just clinical data. To validate the hypothesis, we have combined synthetic clinical data with subset of features from real genomic data. In our results, we observed that prediction accuracy, precision, recall with just genomic data is 80.77%, 20%, 4%. Prediction accuracy, precision, recall with just synthetic clinical data is 93.33%, 95%, 30%. While prediction accuracy, precision, recall for combination of real genomic and synthetic clinical data is 90.83%, 10%, 5%. The combination of real genomic and synthetic clinical data decreased the accuracy since the genomic data is weakly correlated. Thus we conclude that the combination of genomic and clinical data does not improve pancreatic cancer prediction accuracy. A dataset with more significant genomic features might help to predict pancreatic cancer more accurately.