Quantitative Models and Construction Methods 1. Model Name: Genetic Programming Industry Rotation Model - **Model Construction Idea**: This model directly extracts factors from industry indices' price-volume and valuation data, updating the factor library at the end of each quarter. It selects the top five industries with the highest multi-factor composite scores for equal-weight allocation on a weekly basis[3][32] - **Model Construction Process**: - Factor extraction is performed on industry indices based on price-volume and valuation data - The factor library is updated quarterly - Weekly rebalancing is conducted, selecting the top five industries with the highest composite scores for equal-weight allocation[3][32] - **Model Evaluation**: The model achieved strong absolute and relative returns but exhibited rapid industry rotation, leading to slight underperformance against the benchmark in the previous week[3][32] 2. Model Name: Absolute Return ETF Simulated Portfolio - **Model Construction Idea**: Asset allocation weights are determined based on recent trends, with stronger-trending assets assigned higher weights. Equity allocation within the portfolio follows the monthly views of an industry rotation model[4][37] - **Model Construction Process**: - Asset classes are weighted based on recent trend strength - Equity allocation is determined by a monthly industry rotation model - The portfolio includes equity ETFs (e.g., dividend, healthcare, metals) and commodity ETFs (e.g., energy, soybean meal)[4][39] - **Model Evaluation**: The portfolio demonstrated stable performance with a focus on trend-following and diversification[4][37] --- Model Backtesting Results 1. Genetic Programming Industry Rotation Model - **Annualized Return**: 31.87% - **Annualized Volatility**: 18.18% - **Sharpe Ratio**: 1.75 - **Maximum Drawdown**: -19.63% - **Calmar Ratio**: 1.62 - **Year-to-Date (YTD) Return**: 28.68% - **Weekly Performance**: 3.03%[35] 2. Absolute Return ETF Simulated Portfolio - **Annualized Return**: 6.53% - **Annualized Volatility**: 3.82% - **Maximum Drawdown**: -4.65% - **Sharpe Ratio**: 1.71 - **Calmar Ratio**: 1.41 - **Year-to-Date (YTD) Return**: 5.58% - **Weekly Performance**: 0.33%[38] --- Quantitative Factors and Construction Methods 1. Factor Name: Market Intrinsic Momentum Indicators - **Factor Construction Idea**: These indicators measure the internal momentum of the market by analyzing the distribution of individual stock performance[18][19] - **Factor Construction Process**: - **Indicator 1**: Daily turnover difference between rising and falling stocks, normalized by total turnover $ \text{Indicator 1} = \frac{\text{Turnover of rising stocks - Turnover of falling stocks}}{\text{Total turnover}} $ - **Indicator 2**: Monthly high-low turnover difference, normalized by total turnover $ \text{Indicator 2} = \frac{\text{Turnover of stocks hitting monthly highs - Turnover of stocks hitting monthly lows}}{\text{Total turnover}} $ - **Indicator 3**: Six-month high-low turnover difference, normalized by total turnover $ \text{Indicator 3} = \frac{\text{Turnover of stocks hitting six-month highs - Turnover of stocks hitting six-month lows}}{\text{Total turnover}} $ - **Indicator 4**: Annual high-low turnover difference, normalized by total turnover $ \text{Indicator 4} = \frac{\text{Turnover of stocks hitting annual highs - Turnover of stocks hitting annual lows}}{\text{Total turnover}} $[18][19] - **Factor Evaluation**: These indicators effectively capture short-term and long-term market strength and provide strong signals for market trends[19] --- Factor Backtesting Results 1. Market Intrinsic Momentum Indicators - All four indicators showed upward trends in recent periods, aligning with the market's upward trajectory, indicating strong internal momentum supporting the index's rise[19]