Quantitative Models and Construction Methods 1. Model Name: Diffusion Index Model - **Model Construction Idea**: The model is based on the principle of price momentum, aiming to capture upward trends in industries and sectors[22][23] - **Model Construction Process**: The diffusion index is calculated for each industry based on its price momentum. The model ranks industries by their diffusion index values and selects the top-performing industries for portfolio allocation. The model has been tracking out-of-sample performance since 2021, with adjustments made monthly or weekly based on updated diffusion index rankings[22][23] - **Model Evaluation**: The model has shown strong performance in capturing industry trends during momentum-driven markets but struggles during market reversals[22][36] 2. Model Name: GRU Factor Model - **Model Construction Idea**: This model leverages minute-level price and volume data processed through a GRU (Gated Recurrent Unit) deep learning network to generate industry factors for rotation strategies[37] - **Model Construction Process**: The GRU model uses historical price and volume data as input to train a deep learning network. The network identifies patterns and generates factors that are used to rank industries. The top-ranked industries are selected for portfolio allocation. The model is updated weekly to reflect changes in the rankings[30][31][37] - **Model Evaluation**: The GRU model performs well in short-term trading environments but has shown limited effectiveness in long-term scenarios. It is also sensitive to extreme market conditions[37] --- Backtesting Results of Models 1. Diffusion Index Model - **Weekly Average Return**: -5.50% - **Excess Return over Equal-Weighted CSI First-Level Industry Index**: -0.42% - **November-to-Date Excess Return**: -1.13% - **Year-to-Date Excess Return**: 1.22%[26][22][23] 2. GRU Factor Model - **Weekly Average Return**: -4.71% - **Excess Return over Equal-Weighted CSI First-Level Industry Index**: 0.35% - **November-to-Date Excess Return**: 2.92% - **Year-to-Date Excess Return**: -2.74%[35][30][31] --- Quantitative Factors and Construction Methods 1. Factor Name: Diffusion Index - **Factor Construction Idea**: The diffusion index measures the momentum of industries by analyzing price trends and ranks industries based on their momentum[22][23] - **Factor Construction Process**: The diffusion index is calculated for each industry using price momentum data. Industries are ranked based on their diffusion index values, and the top-ranked industries are selected for portfolio allocation. The index is updated weekly or monthly to reflect changes in industry momentum[22][23] - **Factor Evaluation**: The factor effectively captures upward trends in industries but may underperform during market reversals[22][36] 2. Factor Name: GRU Industry Factor - **Factor Construction Idea**: The GRU industry factor is derived from minute-level price and volume data processed through a GRU deep learning network to identify patterns and rank industries[37] - **Factor Construction Process**: The GRU model processes historical price and volume data through a deep learning network. The network generates factors that are used to rank industries. The top-ranked industries are selected for portfolio allocation, with updates made weekly[30][31][37] - **Factor Evaluation**: The factor is effective in short-term trading environments but less so in long-term scenarios. It is also sensitive to extreme market conditions[37] --- Backtesting Results of Factors 1. Diffusion Index Factor - **Weekly Average Return**: -5.50% - **Excess Return over Equal-Weighted CSI First-Level Industry Index**: -0.42% - **November-to-Date Excess Return**: -1.13% - **Year-to-Date Excess Return**: 1.22%[26][22][23] 2. GRU Industry Factor - **Weekly Average Return**: -4.71% - **Excess Return over Equal-Weighted CSI First-Level Industry Index**: 0.35% - **November-to-Date Excess Return**: 2.92% - **Year-to-Date Excess Return**: -2.74%[35][30][31]

Quantitative Models and Construction Diffusion Index Model - **Model Name**: Diffusion Index Model [5][17] - **Construction Idea**: The model monitors the market's diffusion index to identify critical turning points for trading signals [5][17] - **Construction Process**: - The diffusion index is calculated based on the relative price movements of constituent stocks within the micro-cap index over a specific time window [37] - The model uses three methods: - **First Threshold Method (Left-Side Trading)**: Triggered when the diffusion index reaches a predefined risk threshold. For example, on November 14, 2025, the index value of 0.925 triggered a sell signal [41] - **Delayed Threshold Method (Right-Side Trading)**: Provides a sell signal when the index value drops below a delayed threshold, such as 0.8975 on November 17, 2025 [46] - **Dual Moving Average Method (Adaptive Trading)**: Generates buy signals based on the crossover of two moving averages, such as the buy signal on October 13, 2025 [47] - **Evaluation**: The model effectively identifies market turning points and provides actionable trading signals [5][17] Small-Cap Low-Volatility 50 Strategy - **Model Name**: Small-Cap Low-Volatility 50 Strategy [7][16][33] - **Construction Idea**: Selects 50 stocks with small market capitalization and low volatility from the micro-cap index [7][33] - **Construction Process**: - Stocks are screened based on market capitalization and volatility metrics [7][33] - Portfolio is rebalanced bi-weekly [7][33] - Transaction costs are set at 0.3% for both buying and selling [7] - **Evaluation**: The strategy demonstrates strong performance in specific market conditions but underperforms during broader market downturns [7][33] --- Model Backtesting Results Diffusion Index Model - **First Threshold Method**: Triggered sell signal at 0.925 on November 14, 2025 [41] - **Delayed Threshold Method**: Triggered sell signal at 0.8975 on November 17, 2025 [46] - **Dual Moving Average Method**: Generated buy signal on October 13, 2025 [47] Small-Cap Low-Volatility 50 Strategy - **2024 Performance**: Annual return of 7.07%, excess return of -2.93% [7][33] - **2025 YTD Performance**: Annual return of 63.78%, weekly excess return of -2.23% [7][33] --- Quantitative Factors and Construction Weekly Factor Performance - **Top 5 Factors**: - **Leverage Factor**: Weekly rank IC of 0.182, historical average of -0.005 [4] - **Free Float Ratio Factor**: Weekly rank IC of 0.138, historical average of -0.012 [4] - **Turnover Factor**: Weekly rank IC of 0.116, historical average of -0.081 [4] - **Liquidity Factor**: Weekly rank IC of 0.075, historical average of -0.041 [4] - **Dividend Yield Factor**: Weekly rank IC of 0.064, historical average of 0.022 [4] - **Bottom 5 Factors**: - **Unadjusted Stock Price Factor**: Weekly rank IC of -0.311, historical average of -0.017 [4] - **Beta Factor**: Weekly rank IC of -0.3, historical average of 0.003 [4] - **Non-Liquidity Factor**: Weekly rank IC of -0.161, historical average of 0.039 [4] - **Inverse PE_TTM Factor**: Weekly rank IC of -0.138, historical average of 0.016 [4] - **Single-Quarter ROE Factor**: Weekly rank IC of -0.089, historical average of 0.021 [4] Additional Weekly Factor Performance - **Top 5 Factors**: - **Logarithmic Market Cap Factor**: Weekly rank IC of 0.225, historical average of -0.034 [16] - **Nonlinear Market Cap Factor**: Weekly rank IC of 0.225, historical average of -0.034 [16] - **Beta Factor**: Weekly rank IC of 0.083, historical average of 0.003 [16] - **Unadjusted Stock Price Factor**: Weekly rank IC of 0.065, historical average of -0.017 [16] - **Past Year Volatility Factor**: Weekly rank IC of 0.06, historical average of -0.033 [16] - **Bottom 5 Factors**: - **Past 10-Day Return Factor**: Weekly rank IC of -0.226, historical average of -0.061 [16] - **Momentum Factor**: Weekly rank IC of -0.196, historical average of -0.006 [16] - **Leverage Factor**: Weekly rank IC of -0.114, historical average of -0.005 [16] - **Single-Quarter Net Profit Growth Factor**: Weekly rank IC of -0.11, historical average of 0.019 [16] - **Standardized Expected Earnings Factor**: Weekly rank IC of -0.104, historical average of 0.013 [16] --- Factor Backtesting Results Weekly Factor Performance - **Leverage Factor**: Weekly rank IC of 0.182 [4] - **Free Float Ratio Factor**: Weekly rank IC of 0.138 [4] - **Turnover Factor**: Weekly rank IC of 0.116 [4] - **Liquidity Factor**: Weekly rank IC of 0.075 [4] - **Dividend Yield Factor**: Weekly rank IC of 0.064 [4] - **Unadjusted Stock Price Factor**: Weekly rank IC of -0.311 [4] - **Beta Factor**: Weekly rank IC of -0.3 [4] - **Non-Liquidity Factor**: Weekly rank IC of -0.161 [4] - **Inverse PE_TTM Factor**: Weekly rank IC of -0.138 [4] - **Single-Quarter ROE Factor**: Weekly rank IC of -0.089 [4] Additional Weekly Factor Performance - **Logarithmic Market Cap Factor**: Weekly rank IC of 0.225 [16] - **Nonlinear Market Cap Factor**: Weekly rank IC of 0.225 [16] - **Beta Factor**: Weekly rank IC of 0.083 [16] - **Unadjusted Stock Price Factor**: Weekly rank IC of 0.065 [16] - **Past Year Volatility Factor**: Weekly rank IC of 0.06 [16] - **Past 10-Day Return Factor**: Weekly rank IC of -0.226 [16] - **Momentum Factor**: Weekly rank IC of -0.196 [16] - **Leverage Factor**: Weekly rank IC of -0.114 [16] - **Single-Quarter Net Profit Growth Factor**: Weekly rank IC of -0.11 [16] - **Standardized Expected Earnings Factor**: Weekly rank IC of -0.104 [16]

Quantitative Models and Construction Methods - **Model Name**: Diffusion Index Model **Construction Idea**: The model monitors the future critical points of the diffusion index to predict market trends[6][38][39] **Construction Process**: 1. The diffusion index is calculated based on the relative price changes of constituent stocks over a specific time window 2. Horizontal axis represents future price changes (e.g., from +10% to -10%), while vertical axis represents the length of the review window (e.g., 20 days to 10 days) 3. Example: If all constituent stocks drop by 5% after 5 days, the diffusion index value is 0.69[38] 4. The model uses methods like left-side threshold, right-side threshold, and dual moving average to generate signals - Left-side threshold method triggered an opening signal on September 23, 2025, with a value of 0.0575[43] - Right-side threshold method triggered an opening signal on September 25, 2025, with a value of 0.1825[47] - Dual moving average method provided a bullish signal on October 13, 2025[48] **Evaluation**: The model is effective in identifying high-risk zones and generating trading signals[6][39][48] - **Model Name**: Small Cap Low Volatility 50 Strategy **Construction Idea**: Select 50 stocks with small market capitalization and low volatility from micro-cap stocks[8][34] **Construction Process**: 1. Stocks are selected based on market capitalization and volatility criteria 2. Portfolio is rebalanced bi-weekly 3. Transaction cost is set at 0.3% for both sides 4. Benchmark index: Wind Micro-Cap Index (8841431.WI)[8][34] **Evaluation**: The strategy has shown strong performance in 2025, with significant YTD returns[8][34] Model Backtesting Results - **Diffusion Index Model**: - Current diffusion index value: 0.82, indicating a medium-high level[38][39] - Weekly increase from 0.78 to 0.82[39] - Future prediction: If the index rises by 2% next week, it will trigger the risk threshold[39] - **Small Cap Low Volatility 50 Strategy**: - 2024 return: 7.07%, excess return: -2.93%[8][34] - 2025 YTD return: 77.82%, weekly excess return: 1.50%[8][34] Quantitative Factors and Construction Methods - **Factor Name**: Free Float Ratio Factor **Construction Idea**: Measure the proportion of free-floating shares in total shares[5][16][32] **Construction Process**: 1. Calculate the ratio of free-floating shares to total shares 2. Rank IC value for the week: 0.108; historical average: -0.012[5][16][32] **Evaluation**: Positive weekly IC indicates strong predictive power[5][16][32] - **Factor Name**: Leverage Factor **Construction Idea**: Assess the financial leverage of companies[5][16][32] **Construction Process**: 1. Calculate the ratio of total debt to equity 2. Rank IC value for the week: 0.104; historical average: -0.006[5][16][32] **Evaluation**: Positive weekly IC suggests effective factor performance[5][16][32] - **Factor Name**: 10-Day Total Market Cap Turnover Rate Factor **Construction Idea**: Measure the turnover rate of total market capitalization over 10 days[5][16][32] **Construction Process**: 1. Calculate turnover rate as trading volume divided by total market capitalization over 10 days 2. Rank IC value for the week: 0.099; historical average: -0.059[5][16][32] **Evaluation**: Positive weekly IC indicates good predictive ability[5][16][32] - **Factor Name**: 10-Day Free Float Market Cap Turnover Rate Factor **Construction Idea**: Measure the turnover rate of free-floating market capitalization over 10 days[5][16][32] **Construction Process**: 1. Calculate turnover rate as trading volume divided by free-floating market capitalization over 10 days 2. Rank IC value for the week: 0.098; historical average: -0.061[5][16][32] **Evaluation**: Positive weekly IC suggests strong factor performance[5][16][32] - **Factor Name**: Dividend Yield Factor **Construction Idea**: Measure the dividend yield of stocks[5][16][32] **Construction Process**: 1. Calculate dividend yield as annual dividend divided by stock price 2. Rank IC value for the week: 0.065; historical average: 0.022[5][16][32] **Evaluation**: Positive weekly IC indicates reliable factor performance[5][16][32] Factor Backtesting Results - **Free Float Ratio Factor**: Weekly IC: 0.108; historical average: -0.012[5][16][32] - **Leverage Factor**: Weekly IC: 0.104; historical average: -0.006[5][16][32] - **10-Day Total Market Cap Turnover Rate Factor**: Weekly IC: 0.099; historical average: -0.059[5][16][32] - **10-Day Free Float Market Cap Turnover Rate Factor**: Weekly IC: 0.098; historical average: -0.061[5][16][32] - **Dividend Yield Factor**: Weekly IC: 0.065; historical average: 0.022[5][16][32]