This sample set contains data from 235 exploration wells, each well contains six types of attribute information and exploration results, namely structure (ST), favorite structural index (FST), reservoir thickness (RTH), proportion of sandstone content (PS), hydrocarbon generation intensity (HG), cap rock thickness (CRT) and class label. c = 1 stands for productive well, and c = 0 stands for non-productive well. Supervised learning can be performed using the 6 attribute information and class labels, thereby forming an oil and gas risk prediction model