ZHESHANG SECURITIES·2025-02-21 08:07Quantitative Models and Construction Methods 1. Model Name: CSI Artificial Intelligence Theme Index (930713.CSI) - Model Construction Idea: The index selects 50 stocks from the entire market that are involved in providing foundational resources, technologies, and application support for artificial intelligence, reflecting the overall performance of AI-themed listed companies[31][32] - Model Construction Process: - Sample Space: Same as the CSI All Share Index sample space - Selection Method: 1. Filter securities that meet investability criteria and are involved in AI-related fields such as big data, cloud computing, machine learning, and intelligent chips 2. Rank the securities by daily average total market capitalization over the past year and select the top 50 as index samples - Weighting Method: Free-float market capitalization weighting with a tiered adjustment algorithm - Index Calculation: Where adjusted market value = ∑(stock price × adjusted shares × weight factor) - Index Adjustment: - Regular adjustments: Conducted semi-annually in June and December - Temporary adjustments: Made for events like delisting, mergers, or acquisitions[32] - Model Evaluation: The index is characterized by high return elasticity and broad coverage of AI-related industries, making it a representative benchmark for the AI sector[30][31] --- Model Backtesting Results 1. CSI Artificial Intelligence Theme Index - Annualized Return: 20.59%[49] - Annualized Volatility: 35.55%[49] - Maximum Drawdown: -44.99%[49] - Sharpe Ratio: 0.78[49]