Quantitative Models and Construction Methods - **Model Name**: Diffusion Index Model **Model Construction Idea**: Based on price momentum principles, the model identifies upward trends in industries to optimize allocation decisions[23][24][27] **Model Construction Process**: 1. Calculate the diffusion index for each industry based on price momentum 2. Rank industries by their diffusion index values 3. Allocate to industries with the highest diffusion index values **Evaluation**: The model performs well in capturing upward trends but struggles during market reversals or when trends shift to oversold rebounds[23][27] - **Model Name**: GRU Factor Model **Model Construction Idea**: Utilizes GRU (Gated Recurrent Unit) deep learning networks to analyze minute-level volume and price data for industry rotation[31][32][36] **Model Construction Process**: 1. Input minute-level volume and price data into the GRU network 2. Train the model on historical data to identify industry rotation signals 3. Rank industries based on GRU factor scores and allocate accordingly **Evaluation**: The model adapts well to short-term market dynamics but faces challenges in long-term performance and extreme market conditions[31][38] Model Backtesting Results - **Diffusion Index Model**: - Weekly average return: -1.26% - Excess return over equal-weighted industry index: -1.99% - November excess return: -0.74% - Year-to-date excess return: 1.84%[22][27] - **GRU Factor Model**: - Weekly average return: 1.72% - Excess return over equal-weighted industry index: 1.00% - November excess return: 2.69% - Year-to-date excess return: -3.34%[31][36] Quantitative Factors and Construction Methods - **Factor Name**: Diffusion Index **Factor Construction Idea**: Measures industry momentum by tracking price trends and ranking industries accordingly[24][25][26] **Factor Construction Process**: 1. Calculate the diffusion index for each industry using price trend data 2. Rank industries based on diffusion index values 3. Identify industries with the highest and lowest diffusion index values for allocation decisions **Evaluation**: Effective in identifying upward trends but sensitive to market reversals[23][24] - **Factor Name**: GRU Factor **Factor Construction Idea**: Derived from GRU deep learning networks, the factor captures industry rotation signals based on volume and price dynamics[31][32][36] **Factor Construction Process**: 1. Train GRU networks on historical minute-level data 2. Generate GRU factor scores for industries 3. Rank industries by GRU factor scores for allocation decisions **Evaluation**: Strong adaptability to short-term market changes but limited robustness in long-term scenarios[31][38] Factor Backtesting Results - **Diffusion Index Factor**: - Top industries by diffusion index: Nonferrous metals (0.991), Banking (0.968), Steel (0.949), Communication (0.918), Electric equipment & new energy (0.914), Comprehensive (0.885)[24][25][26] - Weekly average return: -1.26% - Excess return over equal-weighted industry index: -1.99% - November excess return: -0.74% - Year-to-date excess return: 1.84%[22][27] - **GRU Factor**: - Top industries by GRU factor: Comprehensive (3.41), Real estate (2.63), Petroleum & petrochemical (2.13), Light manufacturing (1.67), Steel (0.53), Comprehensive finance (0.52)[32][35][36] - Weekly average return: 1.72% - Excess return over equal-weighted industry index: 1.00% - November excess return: 2.69% - Year-to-date excess return: -3.34%[31][36]

- The diffusion index model tracks industry rotation based on momentum principles, focusing on upward trends in industry performance. It has been monitored for four years, with notable performance in 2021 achieving excess returns of over 25% before a significant drawdown in September due to cyclical stock adjustments. In 2025, the model suggests allocating to industries such as non-ferrous metals, banking, communication, steel, electronics, and power equipment & new energy[22][23][26] - The GRU factor model utilizes minute-level volume and price data processed through GRU deep learning networks. It has shown strong adaptability in short cycles but performs less effectively in long cycles. In 2025, the model's industry rotation includes sectors like agriculture, power & utilities, basic chemicals, transportation, steel, and petrochemicals. Weekly average returns were 2.56%, with excess returns of 1.65% against equal-weighted industry benchmarks. Year-to-date excess returns stand at -4.49%[29][30][32] - Diffusion index weekly tracking shows top-ranked industries as non-ferrous metals (0.991), banking (0.931), power equipment & new energy (0.925), communication (0.92), steel (0.871), and electronics (0.864). Industries with the most significant weekly changes include power equipment & new energy (+0.083), petrochemicals (+0.082), and light manufacturing (+0.078)[23][24][25] - GRU factor weekly tracking ranks industries such as comprehensive (7.22), basic chemicals (3.37), building materials (2.7), transportation (2.36), power & utilities (1.96), and food & beverages (1.94) as top performers. Industries with notable weekly increases include power & utilities, non-bank finance, and basic chemicals[30][33][37]

- Model Name: Diffusion Index Model - Model Construction Idea: The model uses the diffusion index to monitor the critical point of future changes in the diffusion index[6][38] - Detailed Construction Process: The model uses the following formula to calculate the diffusion index: $$ \text{Diffusion Index} = \frac{\text{Number of Advancing Stocks}}{\text{Total Number of Stocks}} $$ The model monitors the critical point of future changes in the diffusion index by observing the values of the diffusion index at different time points[38][39] - Model Evaluation: The model is effective in predicting the high volatility of the micro-cap index in the coming week[39] - Testing Results: The current value of the diffusion index is 0.78, indicating a relatively high level[39] - Model Name: Initial Threshold Method (Left-Side Trading) - Model Construction Idea: The model triggers an opening signal when the diffusion index reaches a certain threshold[6][42] - Detailed Construction Process: The model uses the following formula to calculate the threshold: $$ \text{Threshold} = \text{Diffusion Index} \times \text{Historical Average} $$ The model triggered an opening signal on September 23, 2025, when the diffusion index reached 0.0575[42] - Model Evaluation: The model is effective in providing timely trading signals[42] - Testing Results: The model triggered an opening signal on September 23, 2025[42] - Model Name: Delayed Threshold Method (Right-Side Trading) - Model Construction Idea: The model provides an opening signal when the diffusion index reaches a delayed threshold[6][45] - Detailed Construction Process: The model uses the following formula to calculate the delayed threshold: $$ \text{Delayed Threshold} = \text{Diffusion Index} \times \text{Historical Average} + \text{Delay Factor} $$ The model provided an opening signal on September 25, 2025, when the diffusion index reached 0.1825[45] - Model Evaluation: The model is effective in providing delayed but accurate trading signals[45] - Testing Results: The model provided an opening signal on September 25, 2025[45] - Model Name: Dual Moving Average Method (Adaptive Trading) - Model Construction Idea: The model uses dual moving averages to provide trading signals[6][46] - Detailed Construction Process: The model uses the following formula to calculate the dual moving averages: $$ \text{Short-Term Moving Average} = \frac{\sum_{i=1}^{n} \text{Price}_i}{n} $$ $$ \text{Long-Term Moving Average} = \frac{\sum_{i=1}^{m} \text{Price}_i}{m} $$ The model provided a bullish signal on October 13, 2025, when the short-term moving average crossed above the long-term moving average[46] - Model Evaluation: The model is effective in providing adaptive trading signals based on market trends[46] - Testing Results: The model provided a bullish signal on October 13, 2025[46] Factor Construction and Performance - Factor Name: Dividend Yield Factor - Factor Construction Idea: The factor ranks stocks based on their dividend yield[5][16] - Detailed Construction Process: The factor uses the following formula to calculate the dividend yield: $$ \text{Dividend Yield} = \frac{\text{Annual Dividends}}{\text{Stock Price}} $$ The factor ranks stocks from highest to lowest dividend yield[16] - Factor Evaluation: The factor is effective in identifying high-yield stocks[16] - Testing Results: The factor's rank IC for the week is 0.199, with a historical average of 0.022[16] - Factor Name: PB Inverse Factor - Factor Construction Idea: The factor ranks stocks based on the inverse of their price-to-book ratio[5][16] - Detailed Construction Process: The factor uses the following formula to calculate the inverse PB ratio: $$ \text{PB Inverse} = \frac{1}{\text{Price-to-Book Ratio}} $$ The factor ranks stocks from highest to lowest PB inverse[16] - Factor Evaluation: The factor is effective in identifying undervalued stocks[16] - Testing Results: The factor's rank IC for the week is 0.112, with a historical average of 0.034[16] - Factor Name: Illiquidity Factor - Factor Construction Idea: The factor ranks stocks based on their illiquidity[5][16] - Detailed Construction Process: The factor uses the following formula to calculate illiquidity: $$ \text{Illiquidity} = \frac{\text{Absolute Return}}{\text{Trading Volume}} $$ The factor ranks stocks from highest to lowest illiquidity[16] - Factor Evaluation: The factor is effective in identifying illiquid stocks[16] - Testing Results: The factor's rank IC for the week is 0.103, with a historical average of 0.04[16] - Factor Name: Growth Factor - Factor Construction Idea: The factor ranks stocks based on their growth potential[5][16] - Detailed Construction Process: The factor uses the following formula to calculate growth: $$ \text{Growth} = \frac{\text{Current Period Earnings}}{\text{Previous Period Earnings}} - 1 $$ The factor ranks stocks from highest to lowest growth[16] - Factor Evaluation: The factor is effective in identifying high-growth stocks[16] - Testing Results: The factor's rank IC for the week is 0.019, with a historical average of -0.003[16] - Factor Name: Residual Volatility Factor - Factor Construction Idea: The factor ranks stocks based on their residual volatility[5][16] - Detailed Construction Process: The factor uses the following formula to calculate residual volatility: $$ \text{Residual Volatility} = \sqrt{\frac{\sum_{i=1}^{n} (\text{Return}_i - \text{Expected Return})^2}{n}} $$ The factor ranks stocks from highest to lowest residual volatility[16] - Factor Evaluation: The factor is effective in identifying stocks with high residual volatility[16] - Testing Results: The factor's rank IC for the week is 0.015, with a historical average of -0.039[16] Factor Backtesting Results - Dividend Yield Factor: Rank IC for the week is 0.199, historical average is 0.022[16] - PB Inverse Factor: Rank IC for the week is 0.112, historical average is 0.034[16] - Illiquidity Factor: Rank IC for the week is 0.103, historical average is 0.04[16] - Growth Factor: Rank IC for the week is 0.019, historical average is -0.003[16] - Residual Volatility Factor: Rank IC for the week is 0.015, historical average is -0.039[16]

- The diffusion index model has been tracking out-of-sample performance for four years, with notable results in 2021 when momentum strategies captured industry trends, achieving excess returns of over 25% before a significant drawdown in September due to cyclical stock adjustments. In 2022, the strategy maintained stable returns with an annual excess return of 6.12%. However, in 2023, excess returns declined to -4.58%, and in 2024, a major drawdown occurred after September due to the model's focus on upward trends, missing rebound industries, resulting in an annual excess return of -5.82%[24][28] - The diffusion index model suggests allocating to industries such as non-bank finance, construction, and defense military, which showed significant week-on-week improvement in rankings. The top six industries based on diffusion index rankings as of October 24, 2025, are non-bank finance (0.988), banking (0.967), steel (0.952), communication (0.946), comprehensive (0.913), and non-bank finance (0.9)[25][26][27] - The GRU factor model, based on minute-level volume and price data processed through GRU deep learning networks, has shown strong performance in short cycles but weaker performance in long cycles. The model has been effective in capturing trading information since 2021, achieving significant excess returns. However, since February 2025, the model has faced challenges in generating excess returns due to market focus on thematic trading[31][37] - The GRU factor model ranks industries based on their GRU factor scores. As of October 24, 2025, the top six industries are non-bank finance (1.13), banking (1), electric power and utilities (0.54), textile and apparel (0.03), automotive (-0.58), and machinery (-0.73). Industries with the lowest GRU factor scores include food and beverage (-17.79), non-ferrous metals (-10.81), basic chemicals (-8.82), agriculture (-8.76), coal (-6.57), and building materials (-6.48)[6][13][32] - The GRU factor model's weekly industry rotation suggests allocating to non-bank finance, electric power and utilities, textile and apparel, transportation, steel, and petrochemicals. For the week ending October 24, 2025, the model achieved an average return of 1.89%, underperforming the equal-weighted return of the CSI first-tier industries by -0.77%. For October, the model's excess return is 1.80%, while the year-to-date excess return stands at -6.41%[6][34][39]

Quantitative Models and Construction Methods - **Model Name**: Diffusion Index Model **Model Construction Idea**: The model is designed to monitor the future critical points of diffusion index changes and provide trading signals based on different methods such as threshold methods and moving average methods [6][37][38] **Model Construction Process**: - The diffusion index is calculated based on the relative price changes of micro-cap stock index constituent stocks over a specific time window. - The horizontal axis represents the percentage change in stock prices from +10% to -10% over the next N days, while the vertical axis represents the length of the retrospective window T days or the future N days. - For example, at a horizontal axis value of 0.95 and a vertical axis value of 15 days, the diffusion index value is 0.04, indicating that if all micro-cap stock index constituent stocks drop by 5% after N=5 days, the diffusion index value will be 0.04. - The model uses three methods to generate trading signals: - **First Threshold Method (Left-side Trading)**: Triggered a buy signal on September 23, 2025, with a closing value of 0.0575 [40] - **Delayed Threshold Method (Right-side Trading)**: Gave a buy signal on September 25, 2025, with a closing value of 0.1825 [44] - **Double Moving Average Method (Adaptive Trading)**: Provided a sell signal on August 4, 2025 [45] **Model Evaluation**: The diffusion index is currently at a medium level, indicating a short-term downward trend but not expected to trigger the 0.1 threshold in the next 10 trading days [37][38] Model Backtesting Results - **Diffusion Index Model**: - Current diffusion index value: 0.50 [37] - First Threshold Method: Buy signal triggered at 0.0575 on September 23, 2025 [40] - Delayed Threshold Method: Buy signal triggered at 0.1825 on September 25, 2025 [44] - Double Moving Average Method: Sell signal triggered on August 4, 2025 [45] --- Quantitative Factors and Construction Methods - **Factor Name**: Leverage Factor **Factor Construction Idea**: Measures the financial leverage of companies to assess their risk and return potential [5][31] **Factor Construction Process**: - The leverage factor is calculated as the ratio of total debt to total equity. - The rank IC for this factor is calculated weekly to evaluate its predictive power for stock returns [31] **Factor Evaluation**: This factor showed a positive rank IC this week, indicating its effectiveness in predicting stock returns [5][31] - **Factor Name**: Free Float Ratio Factor **Factor Construction Idea**: Evaluates the proportion of shares available for trading in the market [5][31] **Factor Construction Process**: - The free float ratio factor is calculated as the ratio of free float shares to total shares outstanding. - Weekly rank IC is used to measure its predictive ability [31] **Factor Evaluation**: This factor demonstrated a positive rank IC this week, suggesting its utility in forecasting stock performance [5][31] - **Factor Name**: Dividend Yield Factor **Factor Construction Idea**: Measures the dividend yield of stocks to identify value opportunities [5][31] **Factor Construction Process**: - The dividend yield factor is calculated as the ratio of annual dividend per share to the current stock price. - Weekly rank IC is computed to assess its predictive power [31] **Factor Evaluation**: This factor showed a positive rank IC this week, indicating its effectiveness in predicting stock returns [5][31] - **Factor Name**: Single-quarter ROE Factor **Factor Construction Idea**: Measures the return on equity for a single quarter to evaluate profitability [5][31] **Factor Construction Process**: - The single-quarter ROE factor is calculated as the ratio of net income to shareholders' equity for a single quarter. - Weekly rank IC is used to measure its predictive ability [31] **Factor Evaluation**: This factor demonstrated a positive rank IC this week, suggesting its utility in forecasting stock performance [5][31] - **Factor Name**: Growth Factor **Factor Construction Idea**: Measures the growth potential of companies based on financial metrics [5][31] **Factor Construction Process**: - The growth factor is calculated using metrics such as revenue growth, earnings growth, and other growth indicators. - Weekly rank IC is computed to assess its predictive power [31] **Factor Evaluation**: This factor showed a positive rank IC this week, indicating its effectiveness in predicting stock returns [5][31] --- Factor Backtesting Results - **Leverage Factor**: Rank IC this week: 0.176, historical average: -0.006 [5][31] - **Free Float Ratio Factor**: Rank IC this week: 0.156, historical average: -0.013 [5][31] - **Dividend Yield Factor**: Rank IC this week: 0.109, historical average: 0.021 [5][31] - **Single-quarter ROE Factor**: Rank IC this week: 0.091, historical average: 0.022 [5][31] - **Growth Factor**: Rank IC this week: 0.091, historical average: -0.003 [5][31]

- The report introduces the **Diffusion Index Industry Rotation Model**, which tracks industry trends based on momentum strategies. The model has been monitored for four years, with notable performance in 2021 when it captured industry trends effectively, achieving an excess return of over 25% before experiencing a significant drawdown due to cyclical stock adjustments. In 2025, the model suggested allocating to industries such as comprehensive, non-ferrous metals, communication, banking, media, and retail trade[24][28] - The **Diffusion Index Industry Rotation Model** ranks industries weekly based on diffusion index values. As of September 26, 2025, the top six industries were communication (0.949), non-ferrous metals (0.927), banking (0.897), electronics (0.864), automotive (0.859), and comprehensive (0.811). The bottom six industries were food and beverage (0.153), non-bank finance (0.212), coal (0.342), construction (0.348), real estate (0.362), and consumer services (0.415)[25][26][27] - The **GRU Factor Industry Rotation Model** utilizes GRU deep learning networks to analyze minute-level price and volume data. It has shown strong adaptability in short cycles but performs less effectively in long cycles. The model has been operational since 2021, achieving significant excess returns initially. However, in 2025, the model faced challenges in capturing excess returns due to concentrated market themes and speculative trading[31][37] - The **GRU Factor Industry Rotation Model** ranks industries weekly based on GRU factor values. As of September 26, 2025, the top six industries were steel (3.15), real estate (2.6), building materials (2.08), petroleum and petrochemicals (1.85), transportation (0.81), and electric power and utilities (0.01). The bottom six industries were computing (-32.91), media (-29.46), communication (-17.57), food and beverage (-13.4), pharmaceuticals (-13.36), and non-ferrous metals (-12.73)[6][13][32] - The **Diffusion Index Industry Rotation Model** achieved an average weekly return of -0.00%, with an excess return of 0.78% compared to the equal-weighted return of CICC primary industries. Since September, the model has recorded an excess return of -1.10%, and a year-to-date excess return of 3.68%[23][28] - The **GRU Factor Industry Rotation Model** recorded an average weekly return of -0.61%, with an excess return of 0.17% compared to the equal-weighted return of CICC primary industries. Since September, the model has achieved an excess return of 0.07%, and a year-to-date excess return of -7.53%[31][34]

Quantitative Models and Construction Methods 1. Model Name: Diffusion Index Industry Rotation Model - **Model Construction Idea**: The model is based on the principle of price momentum, aiming to capture upward trends in industries through a diffusion index[26][27] - **Model Construction Process**: 1. Calculate the diffusion index for each industry based on price momentum 2. Rank industries by their diffusion index values 3. Select top industries for allocation based on their rankings 4. Adjust the portfolio monthly or weekly based on updated diffusion index rankings[26][27] - **Model Evaluation**: The model has shown stable performance in certain years (e.g., 2022 with an annual excess return of 6.12%) but struggled during market reversals or concentrated market themes, such as in 2024 and 2025[26][33] 2. Model Name: GRU Factor Industry Rotation Model - **Model Construction Idea**: This model leverages GRU (Gated Recurrent Unit) deep learning networks to process high-frequency volume and price data, aiming to identify industry rotation opportunities[38] - **Model Construction Process**: 1. Input high-frequency volume and price data into the GRU network 2. Train the GRU model on historical data to identify patterns in industry rotation 3. Generate factor scores for industries based on the GRU model's output 4. Rank industries by their GRU factor scores and allocate to top-ranked industries[38][34] - **Model Evaluation**: The model performs well in short cycles but struggles in long cycles or extreme market conditions. It has shown difficulty in capturing excess returns in concentrated market themes during 2025[33][38] --- Model Backtesting Results 1. Diffusion Index Industry Rotation Model - **Weekly Average Return**: -1.74%[30] - **Excess Return (Weekly)**: -1.41%[30] - **Excess Return (September 2025)**: -1.88%[30] - **Excess Return (2025 YTD)**: 2.76%[25][30] 2. GRU Factor Industry Rotation Model - **Weekly Average Return**: -0.72%[36] - **Excess Return (Weekly)**: -0.38%[36] - **Excess Return (September 2025)**: -0.10%[36] - **Excess Return (2025 YTD)**: -7.78%[33][36] --- Quantitative Factors and Construction Methods 1. Factor Name: Diffusion Index - **Factor Construction Idea**: Measures the breadth of price momentum across industries to identify upward trends[26][27] - **Factor Construction Process**: 1. Calculate the proportion of stocks in an industry with positive price momentum 2. Aggregate these proportions to derive the diffusion index for the industry 3. Rank industries based on their diffusion index values[27][28] - **Factor Evaluation**: Effective in capturing upward trends but vulnerable to reversals and underperformance in counter-trend markets[26][33] 2. Factor Name: GRU Factor - **Factor Construction Idea**: Utilizes GRU deep learning to analyze high-frequency trading data and generate predictive scores for industry rotation[38] - **Factor Construction Process**: 1. Input high-frequency trading data into the GRU network 2. Train the model to recognize patterns in industry rotation 3. Output factor scores for industries based on the model's predictions[38][34] - **Factor Evaluation**: Strong in short-term predictions but less effective in long-term or extreme market conditions[33][38] --- Factor Backtesting Results 1. Diffusion Index - **Top Industries (Weekly)**: Non-ferrous Metals (0.978), Banking (0.968), Communication (0.946), Electronics (0.877), Automotive (0.874), Retail (0.873)[27] - **Bottom Industries (Weekly)**: Food & Beverage (0.354), Real Estate (0.46), Coal (0.487), Transportation (0.543), Construction (0.574), Building Materials (0.618)[27] 2. GRU Factor - **Top Industries (Weekly)**: Non-ferrous Metals (7.4), Petrochemicals (5.38), Coal (4.17), Steel (4.15), Building Materials (3.46), Non-banking Financials (3.08)[34] - **Bottom Industries (Weekly)**: Comprehensive Finance (-19.42), Utilities (-13.41), Electronics (-13.18), Pharmaceuticals (-11.14), Automotive (-10.07), Consumer Services (-10.04)[34]

- Model Name: Diffusion Index Model; Construction Idea: The model is based on the observation of industry diffusion indices to capture industry trends; Construction Process: The model tracks the weekly and monthly changes in diffusion indices for various industries, ranking them based on their diffusion index values. The formula used is not explicitly mentioned, but the ranking is based on the diffusion index values observed; Evaluation: The model has shown varying performance over the years, with notable returns in some years and significant drawdowns in others[4][24][25] - Model Name: GRU Factor Model; Construction Idea: The model utilizes GRU (Gated Recurrent Unit) neural networks to process minute-level volume and price data to generate industry factors; Construction Process: The model ranks industries based on GRU-generated factors, which are derived from deep learning on historical volume and price data. The specific formula is not provided, but the ranking is based on the GRU factor values; Evaluation: The model has shown strong performance in short cycles but struggles in longer cycles and extreme market conditions[5][30][31] Model Backtest Results - Diffusion Index Model, Average Weekly Return: 3.95%, Excess Return over Equal-weighted Index: 1.94%, August Excess Return: 1.51%, Year-to-date Excess Return: 1.75%[28] - GRU Factor Model, Average Weekly Return: -0.06%, Excess Return over Equal-weighted Index: -2.07%, August Excess Return: -1.78%, Year-to-date Excess Return: -6.66%[33] Factor Construction and Evaluation - Factor Name: Diffusion Index; Construction Idea: The factor is constructed by observing the weekly and monthly changes in industry diffusion indices; Construction Process: The factor ranks industries based on their diffusion index values, with higher values indicating stronger trends. The specific formula is not provided, but the ranking is based on the observed diffusion index values; Evaluation: The factor has shown varying performance, capturing industry trends effectively in some periods while underperforming in others[4][24][25] - Factor Name: GRU Industry Factor; Construction Idea: The factor is generated using GRU neural networks to process minute-level volume and price data; Construction Process: The factor ranks industries based on GRU-generated values, which are derived from deep learning on historical data. The specific formula is not provided, but the ranking is based on the GRU factor values; Evaluation: The factor performs well in short cycles but faces challenges in longer cycles and extreme market conditions[5][30][31] Factor Backtest Results - Diffusion Index Factor, Top Industries: Comprehensive Finance (1.0), Steel (1.0), Non-bank Finance (0.999), Comprehensive (0.998), Non-ferrous Metals (0.997), Communication (0.997)[25] - GRU Industry Factor, Top Industries: Non-ferrous Metals (5.67), Non-bank Finance (4.65), Building Materials (4.14), Real Estate (4.08), Steel (3.64), Basic Chemicals (2.71)[31][13]

- Model Name: Diffusion Index Model; Model Construction Idea: The model is based on the principle of price momentum; Model Construction Process: The model tracks the weekly and monthly changes in the diffusion index of various industries, ranking them accordingly. The formula used is $ \text{Diffusion Index} = \frac{\text{Number of Upward Trends}}{\text{Total Number of Trends}} $; Model Evaluation: The model has shown varying performance over the years, with significant returns in some periods and notable drawdowns in others[27][28][31] - Model Name: GRU Factor Model; Model Construction Idea: The model utilizes GRU deep learning networks to analyze minute-level volume and price data; Model Construction Process: The model ranks industries based on GRU factors, which are derived from deep learning algorithms processing historical trading data. The formula used is $ \text{GRU Factor} = \text{GRU Network Output} $; Model Evaluation: The model performs well in short cycles but has mixed results in longer cycles[33][34][36] - Diffusion Index Model, Average Weekly Return: 2.06%, Excess Return: -0.00%, August Excess Return: -0.45%, Year-to-Date Excess Return: -0.41%[31] - GRU Factor Model, Average Weekly Return: 2.71%, Excess Return: 0.65%, August Excess Return: 0.32%, Year-to-Date Excess Return: -4.35%[36] - Factor Name: GRU Industry Factor; Factor Construction Idea: The factor is derived from GRU deep learning networks analyzing minute-level trading data; Factor Construction Process: The factor ranks industries based on GRU network outputs, which are calculated from historical volume and price data. The formula used is $ \text{GRU Factor} = \text{GRU Network Output} $; Factor Evaluation: The factor has shown significant changes in rankings, indicating its sensitivity to market conditions[6][14][34] - GRU Industry Factor, Steel: 2.82, Building Materials: 1.72, Transportation: 1.3, Oil & Petrochemicals: 0.27, Construction: -0.46, Comprehensive: -1.87[6][14][34]

- Model Name: Diffusion Index Model; Construction Idea: The model is based on the principle of price momentum, capturing industry trends through diffusion indices; Construction Process: The model tracks the weekly and monthly changes in the diffusion indices of various industries, ranking them accordingly. The formula for the diffusion index is not explicitly provided; Evaluation: The model has shown varying performance over the years, with significant drawdowns during market reversals[24][25][28] - Model Name: GRU Factor Model; Construction Idea: The model utilizes GRU (Gated Recurrent Unit) deep learning networks to process minute-level volume and price data, aiming to capture trading information; Construction Process: The model ranks industries based on GRU factors, which are derived from the deep learning network's analysis of trading data. The specific formula for GRU factors is not provided; Evaluation: The model has performed well in short cycles but has shown general performance in longer cycles[31][32][35] - Diffusion Index Model, Average Weekly Return: 0.89%, Excess Return Since July: -3.47%, Excess Return YTD: -0.45%[28] - GRU Factor Model, Average Weekly Return: 4.27%, Excess Return Since July: 1.34%, Excess Return YTD: -4.25%[35] - Factor Name: Diffusion Index; Construction Idea: The factor is based on the momentum of industry prices, capturing upward trends; Construction Process: The factor is calculated by observing the weekly and monthly changes in the diffusion indices of various industries. The specific formula is not provided; Evaluation: The factor has shown varying performance, with significant drawdowns during market reversals[24][25][28] - Factor Name: GRU Factor; Construction Idea: The factor is derived from GRU deep learning networks, capturing trading information from minute-level volume and price data; Construction Process: The factor is calculated by ranking industries based on the GRU network's analysis of trading data. The specific formula is not provided; Evaluation: The factor has performed well in short cycles but has shown general performance in longer cycles[31][32][35] - Diffusion Index Factor, Top Industries: Comprehensive Finance (1.0), Steel (1.0), Non-Bank Finance (0.999), Comprehensive (0.998), Non-Ferrous Metals (0.997), Home Appliances (0.995)[25] - GRU Factor, Top Industries: Banking (3.3), Real Estate (0.58), Oil & Petrochemicals (-1.26), Textile & Apparel (-1.73), Light Manufacturing (-2.49), Electric Power & Utilities (-2.83)[32]