Guidelines to Select Machine Learning Scheme for Classification of Biomedical Datasets
EvoBIO '09 Proceedings of the 7th European Conference on Evolutionary Computation, Machine Learning and Data Mining in Bioinformatics
Machine learning of clinical performance in a pancreatic cancer database
Artificial Intelligence in Medicine
| Hi-index | 0.00 |
Our goal in this research is to construct predictive models for clinicalperformance of pancreatic cancer patients.Current predictive model design in medical oncologyliterature is dominated by linear and logistic regressiontechniques. We seek to show that novel machine learning methods canperform as well or better than these traditional techniques.We construct these predictive models via a clinical database we havedeveloped for the University of Massachusetts Memorial Hospitalin Worcester, Massachusetts, USA. The database contains retrospectiverecords of 91 patient treatments for pancreatic tumors.Classification and regression predictiontargets include patient survival time, ECOG quality of life scores, surgical outcomes,and tumor characteristics. The predictive accuracy of various data miningmodels is described, and specific models are presented.