A-Share Direction Classification with Technical Indicators and Model Voting

Summary

This shared open-source example builds a daily classifier for Chinese stocks using price and volume data, technical indicators, and three supervised models: a random forest, logistic regression, and a linear support vector classifier. Features include Bollinger Bands, MACD, KDJ, CCI, RSI, Williams %R, cross signals, and overbought or oversold flags. The models classify whether prices will be higher over several forward horizons, and a majority vote turns their predictions into an up-or-down label.

The document includes indicator calculations and target-label construction, plus a process for grouping historical data by stock and exporting predictions. It says the original competition solution used free Tushare data and was shared from an external source; it does not present validation scores or trading results. The excerpt also leaves out portions of the implementation, and its target construction and use of the latest row for fitting and prediction require careful review before results can be trusted. It describes a modeling example, not an established profitable strategy.

Key ideas

• The example combines technical indicators with three classification algorithms.
• Its features include band levels, oscillator values, indicator crosses, and threshold flags.
• Separate labels represent price direction over multiple forward horizons.
• The three model outputs are combined through a majority-vote rule.
• No validation metrics or live trading results are reported.

Tags

This summary was written by Stratmill's research agent from the original; it is not a copy of the source.