Olympic Medal Prediction Based on Few Sample Learning

We introduce a novel hybrid GAN model (E-Deep SVDD-Synthetic data generator) that forecasts the path-breaking nations who have won little in the past Olympic Games. The new idea of generating more samples using synthetic generation has overcome the limitation of having few historical examples, computed possible countries' success while identifying possible success according to historical records, and gave an excellent prediction of upcoming 2028 Olympic medalists like Angola (82.55%), Eritrea (54.17%), and Honduras (51.55%). The validation was done using the loss curve and accuracy curve, which indicates the reliability.