https://tbrain.trendmicro.com.tw/Competitions/Details/2
preprocess.py
feature_select.py
train_model.py
Rank: 4/487
國外各式股市指數漲跌(From: http://finance.yahoo.com/ )、外匯漲跌(From: http://www.taifex.com.tw/chinese/3/3_5.asp )、與 ETF 關聯性最高的前100名個股漲跌。取25天。
用Boruta做特徵選取, cv 找最佳參數,以 Gradient Boosting 的方式預測各個 ETF 各天的漲跌幅。