Hello! This is a stock selection model designed to recommend the top-performing stocks to investors on WeChat Mini Program.
For the recommendation algorithms:
- Conducted data cleaning & exploratory data analysis on 4GB+ stock factor data.
- Labeled stocks, performed feature engineering and selected important features using IC value calculations.
- Built XGBoost, Random Forest, SVM models to predict stock performance, applied Soft Voting to ensemble the 3 models, achieved 0.86 ROC-AUC and successfully selected top-performing stocks.