GLD Machine Learning Strategy (Lasso Feature Selection) by hiron2
By hiron2
Performance Metrics
- Author: hiron2
- Symbol: AMEX:GLD
- Timeframe: 1 day
- Net P&L: +410.21 USD (+0.04%)
- Win Rate: 71.4%
- Profit Factor: 6.488
- Max Drawdown: 77.56 USD (0.01%)
- Total Trades: 14
Description
This script implements a GLD trading strategy derived from the paper:An Extension of Temporal Feature Selectionzenodo.org/records/18729703The strategy is based on a sparse linear model selected by Lasso from moving-average-based features constructed from OHLCV data. The underlying research extends Temporal Feature Selection (TFS) to include polynomial and interaction terms of moving averages, and evaluates the resulting signals on financial time series.This TradingView version reproduces the GLD strategy corresponding to the best Sharpe-ratio result reported in the study.Instrument: GLDDesigned timeframe: Daily (1D)Important notes:- The model is instrument-specific and was selected for GLD.- Performance may differ substantially on other symbols or timeframes.- The result reported in the paper depends on the historical sample, execution assumptions, and backtest settings.- This script is provided for research and educational purposes only and is not investment advice.Paper:zenodo.org/records/18729703