主动量化周报：双线作战：配置看消费，交易在科创

Quantitative Models and Construction Methods 1. Model Name: Fund Position Monitoring Model - Model Construction Idea: The model estimates the allocation of active equity funds across various industries to identify underweight or overweight sectors, providing insights into potential investment opportunities[3][11]. - Model Construction Process: The model uses statistical methods to estimate the proportion of active equity funds allocated to different industries. For example, as of December 20, the allocation to the food and beverage industry was 5.1%, compared to 5.5% in the Wind All A Index, indicating an underweight position. Similarly, allocations to retail and textile industries were approximately neutral[11]. - Model Evaluation: The model provides a clear view of fund allocation trends, helping to identify sectors with potential for increased institutional investment[11]. 2. Model Name: AI Price-Volume Model - Model Construction Idea: This model leverages daily price and volume data to extract trend, volatility, and price-volume characteristics of industry indices. It uses deep neural networks to predict the configuration value of industries[16][18]. - Model Construction Process: The model standardizes industry index scores to a range of [-1,1]. For the current period, the top industries identified by the model include automobiles, machinery, transportation, media, and banking. The previous period's top industries were machinery, transportation, banking, automobiles, and non-bank finance[18]. - Model Evaluation: The model effectively identifies short-term industry opportunities, providing actionable insights for sector allocation[16][18]. 3. Model Name: GDPNOW Model - Model Construction Idea: This model predicts GDP growth rates using high-frequency macroeconomic data, offering real-time insights into economic trends[25]. - Model Construction Process: The model's latest prediction for Q4 2024 GDP growth is 4.6%, unchanged from the previous week. The model updates its forecasts based on new macroeconomic data, ensuring timely and accurate predictions[25][26]. - Model Evaluation: The model provides stable and reliable GDP growth forecasts, aiding in macroeconomic analysis and decision-making[25]. 4. Model Name: Informed Trader Activity Indicator - Model Construction Idea: This indicator measures the activity level of informed traders to gauge market sentiment and predict short-term market trends[27]. - Model Construction Process: The indicator showed a slight decline followed by a rebound during the week, suggesting that informed traders remain optimistic about the market's short-term outlook[27][28]. - Model Evaluation: The indicator offers valuable insights into market sentiment, complementing other timing strategies[27]. --- Model Backtesting Results 1. Fund Position Monitoring Model - Key Metrics: - Food and Beverage Allocation: 5.1% (vs. 5.5% in Wind All A Index)[11] - Retail and Textile Allocation: Neutral[11] 2. AI Price-Volume Model - Key Metrics: - Current Top Industries: Automobiles (0.42), Machinery (0.40), Transportation (0.35), Media (0.32), Banking (0.25)[18][19] - Previous Top Industries: Machinery, Transportation, Banking, Automobiles, Non-Bank Finance[18] 3. GDPNOW Model - Key Metrics: - Q4 2024 GDP Growth Prediction: 4.6%[25][26] 4. Informed Trader Activity Indicator - Key Metrics: - Recent Activity Trend: Decline followed by rebound, indicating optimism[27][28] --- Quantitative Factors and Construction Methods 1. Factor Name: BARRA Style Factors - Factor Construction Idea: These factors measure various stock characteristics, such as value, growth, momentum, and volatility, to identify market preferences and trends[37]. - Factor Construction Process: The factors are calculated based on stock-level data. For example, this week's factor performance includes: - Momentum: 0.3% - Nonlinear Size: 0.3% - Size: 0.3% - Dividend Yield: 0.2% - Earnings Quality: 0.1%[37][38] - Factor Evaluation: The factors provide a comprehensive view of market preferences, aiding in portfolio construction and risk management[37]. --- Factor Backtesting Results 1. BARRA Style Factors - Key Metrics: - Momentum: 0.3% - Nonlinear Size: 0.3% - Size: 0.3% - Dividend Yield: 0.2% - Earnings Quality: 0.1%[37][38]