证券研究报告 | 金融工程 gszqdatemark 2025 09 21 年 月 日 量化周报 非银离确认日线级别下跌仅有一步之遥 非银离确认日线级别下跌仅有一步之遥。本周（9.15-9.19），大盘震荡下 行，上证指数全周收跌 1.30%。在此背景下，大部分板块迎来了一波 30 分钟级别调整，非银离确认日线级别下跌仅有一步之遥。市场的本轮上涨 自 4 月 7 日以来，日线级别反弹已经持续了 5 个多月，反弹幅度也基本在 30%左右，各大指数和板块的上涨基本都轮动了一遍，超 2/3 的行业日线 级别上涨处于超涨状态，几乎所有的规模指数及一半以上的行业更是走出 了复杂的 9-15 浪的上涨结构，而银行也已经率先形成了日线级别下跌， 非银离确认日线级别下跌仅有一步之遥。因此我们认为本轮日线级别上涨 大概率已临近尾声。短期，市场的波动进一步加大后，投资者后续可积极 关注市场未来是否出现放量滞涨、放量大跌及缩量反弹迹象。中期来看， 上证指数、上证 50、沪深 300、中证 500、深证成指、创业板指、科创 50 纷纷确认周线级别上涨，而且在日线上只走出了 3 浪结构，中期牛市刚刚 开始；此外，已有 26 个行业处于 ...

量化周报 证券研究报告 | 金融工程 gszqdatemark 2025 09 07 年 月 日 市场波动进一步加大 市场波动进一步加大。本周（9.1-9.5），大盘出现大幅震荡，上证指数全 周收跌 1.18%。市场的本轮上涨自 4 月 7 日以来，日线级别反弹已经持续 了 4 个多月，反弹幅度也基本在 20%以上，各大指数和板块的上涨基本 都轮动了一遍，上证指数、上证 50、非银、有色、农林牧渔、消费者服务、 商贸零售、纺织服装、计算机、建筑、交运等板块更是走出了复杂的上涨 结构，而银行也已经率先形成了日线级别下跌，因此我们认为本轮日线级 别上涨大概率已临近尾声。短期，市场的波动进一步加大，投资者后续可 积极关注市场未来是否出现放量滞涨、放量大跌及缩量反弹迹象。中期来 看，上证指数、上证 50、沪深 300、中证 500、深证成指、创业板指、科 创 50 纷纷确认周线级别上涨，而且在日线上只走出了 3 浪结构，中期牛 市刚刚开始；此外，已有 25 个行业处于周线级别上涨中，且 17 个行业周 线上涨走了 1-3 浪结构，因此我们认为本轮牛市是个普涨格局。中期对于 投资者而言，仍然可以逆势布局。 A 股景气指 ...

Quantitative Models and Construction Methods - **Model Name**: Index Enhanced Portfolio **Construction Idea**: The model aims to outperform benchmark indices by leveraging quantitative strategies and factor exposures[2][48] **Construction Process**: - The portfolio is constructed using a strategy model that selects stocks based on factor exposures and optimization techniques - The model incorporates historical data and factor analysis to identify stocks with high expected returns relative to the benchmark - Portfolio weights are optimized to maximize excess returns while controlling for risk and tracking error[48][49][54] **Evaluation**: The model has demonstrated consistent excess returns over its benchmark indices, showcasing its effectiveness in active management[48][54] - **Model Name**: Factor Attribution Model **Construction Idea**: This model decomposes portfolio or index returns into contributions from various style factors to understand performance drivers[68] **Construction Process**: - The model uses the BARRA factor framework, which includes factors such as size (SIZE), beta (BETA), momentum (MOM), residual volatility (RESVOL), non-linear size (NLSIZE), valuation (BTOP), liquidity (LIQUIDITY), earnings yield (EARNINGS_YIELD), growth (GROWTH), and leverage (LVRG)[58][68] - Factor exposures are calculated for each stock in the portfolio or index - Portfolio returns are attributed to factor contributions using regression-based methods[68] **Evaluation**: The model provides valuable insights into the sources of portfolio performance, aiding in strategy refinement and risk management[68] Model Backtesting Results - **Index Enhanced Portfolio**: - **Mid-Cap Enhanced Portfolio (CSI 500)**: Weekly return of 1.99%, outperforming the benchmark by 0.22%; cumulative excess return since 2020: 50.26%; maximum drawdown: -4.99%[48][49] - **Large-Cap Enhanced Portfolio (CSI 300)**: Weekly return of 1.85%, outperforming the benchmark by 0.61%; cumulative excess return since 2020: 34.90%; maximum drawdown: -5.86%[54][56] Quantitative Factors and Construction Methods - **Factor Name**: Momentum (MOM) **Construction Idea**: Captures the tendency of stocks with strong past performance to continue outperforming in the short term[58][59] **Construction Process**: - Momentum is calculated as the cumulative return over a specified look-back period (e.g., 6 months or 12 months) - Stocks are ranked based on their momentum scores, and portfolios are constructed by overweighting high-momentum stocks[58][59] **Evaluation**: Momentum factor exhibited high excess returns during the week, indicating strong market preference for trending stocks[59] - **Factor Name**: Beta (BETA) **Construction Idea**: Measures the sensitivity of a stock's returns to market movements, capturing risk exposure[58][59] **Construction Process**: - Beta is calculated using regression analysis of stock returns against market returns over a historical period - High-beta stocks are identified and analyzed for their risk-return trade-offs[58][59] **Evaluation**: High-beta stocks performed well during the week, reflecting market preference for riskier assets[59] - **Factor Name**: Growth (GROWTH) **Construction Idea**: Represents the expected earnings growth of a company, capturing future potential[58][59] **Construction Process**: - Growth is estimated using forward-looking metrics such as analyst earnings forecasts and historical growth rates - Stocks are ranked based on growth scores, and portfolios are constructed by overweighting high-growth stocks[58][59] **Evaluation**: Growth factor underperformed during the week, indicating reduced market preference for growth-oriented stocks[59] Factor Backtesting Results - **Momentum Factor**: Weekly excess return was significantly positive, outperforming other style factors[59][66] - **Beta Factor**: High-beta stocks showed strong performance, contributing positively to portfolio returns[59][66] - **Growth Factor**: Underperformed during the week, reflecting weak market sentiment toward growth stocks[59][66] Additional Observations - **Sector Factors**: - Defense, metals, and coal sectors exhibited high excess returns relative to market-cap-weighted benchmarks[59][63] - Sectors such as healthcare, IT, and media experienced significant drawdowns[59][63] - **Market Sentiment**: - Sentiment indicators based on volatility and trading volume suggest a bullish outlook for the market[36][39][41]