paperswithbacktest/Stocks-Weekly-PharmaClinicalPredict

该数据集提供了全球制药赞助商的每周临床试验结果预测，由SOV.AI提供。每行数据代表在给定周内对一项试验的最新模型视图，结合了监管成功概率、完成时间预期和经济影响指数。数据集覆盖了制药和生物技术试验（带有映射的公司股票代码，如果可用），每周更新（每周评估约1,052项新试验），模型性能在保留验证数据上达到87% ROC-AUC。输出包括成功概率、综合成功分数、预测持续时间（天）、经济效应指数和赞助商类别。

Weekly clinical trial outcome predictions for global pharmaceutical sponsors, provided by SOV.AI. Each row represents the latest modelled view for a trial on a given week, combining regulatory success probabilities, time-to-completion expectations, and an economic impact index. Coverage includes pharmaceutical and biotech trials with mapped corporate tickers (where available), update cadence is weekly (~1,052 new trials evaluated each week), and performance is 87% ROC-AUC on held-out validation data. Outputs include success probability, composite success score, predicted duration (days), economic effect index, and sponsor class.