Estimating the Survivability of Patients Using Prediction Markets and Cox Hazard Proportion Regression

Henry Asante Antwi, Zhou Lulin, Numair Nisar


Available literature shows that a large group's aggregated answers to questions involving quantity estimation, general world knowledge, and spatial reasoning has generally been found to be as good as, and often better than, the answer given by any of the individuals within the group. This has inspired the development of new models of crowd wisdom techniques such as prediction markets. To find out whether prediction market or crowd wisdom tools can help in making accurate prognosis, one medical case was chosen which is how well experienced doctors can predict the survivability of patients with gastric cancer. The results are compared with the outcome of survivability forecast using the Cox hazard proportion regression model and the artificial neural networks. The ANN accurately forecasted 31% of patients to survive whiles 33% will not. On the other hand, the Cox Hazard Model accurately predicted 29% of the patients to survive whiles 31% will not. Finally, the PM market predicted 31% of the patients to survive whiles 31 percent will not survive. On the whole the prediction accuracy of the ANN was 64% whiles that of the CPH and the Prediction Market were 60% and 62% respectively. This implies that that whiles the ANN defeated the PM Model in predicting accuracy of survivability of patients with gastric cancer, it outperformed the Cox Hazard Model by 32 percentage points.

Full Text:



  • There are currently no refbacks.