登录新浪财经APP 搜索【信披】查看更多考评等级 重要提示 7、投资人在首次认购本基金时，如尚未开立苏新基金管理有限公司基金账户，需按销售机构的规定， 提出开立苏新基金管理有限公司基金账户和销售机构交易账户的申请。开户和认购申请可同时办理，一 次性完成，但认购申请的确认须以开户确认成功为前提条件。一个投资人只能开立和使用一个基金账 户，已经开立苏新基金管理有限公司基金账户的投资人可免予开户申请。 1、苏新睿见量化选股股票型证券投资基金（以下简称"本基金"）的募集申请已于2025年8月15日获中国 证监会证监许可〔2025〕1774号文准予募集注册。中国证监会对本基金募集的注册，并不表明其对本基 金的投资价值、市场前景和收益做出实质性判断或保证，也不表明投资于本基金没有风险。 2、本基金为契约型开放式证券投资基金。 3、本基金根据认购费、申购费和销售服务费等费率收取方式的不同，将基金份额分为A、C两类份额。 在投资者认购、申购基金时收取认购、申购费用、赎回时收取赎回费用，并不再从本类别基金资产中计 提销售服务费的，称为A类基金份额；从本类别基金资产中计提销售服务费、赎回时收取赎回费用，而 不收取认购、申购费用的 ...

2025年9月26日，中信期货2025年秋季策略会圆满收官。本次策略会以"潮汐涌动.破卷立新"为主题，采 用线上会议的形式举行。通过7大论坛，中信期货研究所的研究员们以多元视角解析市场，多维度深挖 投资机遇，针对四季度及2026年大宗商品、权益、汇率等期货及衍生品领域展开深入剖析，助力投资者 多维掘金、稳健布局。 宏观与贵金属论坛：稳中求进|宏观趋势与资产配置 中信期货研究所宏观与国际研究部总经理姜婧女士表示，四季度宏观和大类资产基调是"稳中求进"，政 策与基本面对冲效应使得宏观环境偏"稳"，资产配置策略偏"进"。从国内层面看，四季度稳增长政策聚 焦三大方向：一是5000亿元政策性金融工具，二是货币政策大概率实施降准，三是"十五五"规划前瞻。 从海外维度看，美国经济动能趋疲，但政策"再校准"将提供支撑，未来1-2个季度，全球宽松流动性与 财政杠杆带动的复苏预期，将为风险资产提供支撑。 中信期货研究所宏观研究组资深研究员朱善颖女士聚焦贵金属，强调四季度黄金震荡偏强，是长期战略 配置窗口。尽管近两年实际利率与金价相关性降低，但受降息周期重启和美联储独立性风险影响，美国 实际利率下行周期中黄金上涨从未缺席。长期来看 ...

在公募量化产品线中，除了指数增强型品种外，量化选股型品种也占据重要地位，截止 2025 年二季末，全市场 共计 277 只量化选股型基金，合计管理规模 903.20 亿元。这类基金由于不属于标准指数增强范畴，因而，不受限于 指数成分股的投资比例和跟踪误差的硬性约束，拥有更广泛的投资范围和更高的风格暴露自由度，其业绩往往也展现 出更高的弹性。同时，得益于相对宽松的策略环境，不同基金经理能够根据自身的偏好及特长，构建更具差异化的投 资策略。 然而，对于投资者而言，由于量化选股型基金的工具化属性不如指数增强型基金明确，在产品选择时，往往面临 对其定位与策略认知不够清晰的挑战。为此，本篇专题将对部分量化选股型基金的投资策略框架进行梳理，主要聚焦 信达澳亚基金冯玺祥、国泰基金高崇南、信达澳亚基金林景艺、鹏华基金时赟超、西部利得基金翟梓舰等 5 名在量化 选股型产品上投资框架体系各有特色的基金经理（按照姓名拼音字母顺序，下同），涵盖风格配置、中小市值、偏股 混增强等差异化产品定位，以供投资者参考。 冯玺祥（信达澳亚基金）：采用统一框架管理，十分看重因子在整个投资域的有效性以及阿尔法模型的普适性， 通过多样化的因子和模型提 ...

Core Viewpoint - The financing balance of the two markets has surpassed 2.3 trillion yuan, marking a historical high since 2015, indicating a significant increase in liquidity and investor risk appetite during the current bull market [2][3]. Group 1: Exposure of Micro-Cap Stocks - There is a noticeable differentiation in the exposure of quantitative long products to micro-cap stocks this year, with micro-cap indices significantly outperforming mid and large-cap stocks [4][5]. - The weighted discount rate of IC/IM stock index futures has remained high, suggesting an increased exposure of quantitative managers to micro-cap stocks [7]. - In the first quarter, the proportion of holdings in stocks below the 2000 index was about 20-40%, which may rise to over 50% in the third quarter [8]. Group 2: Reasons and Risks of Exposure to Micro-Cap Stocks - Historically, small-cap stocks have shown higher average annualized beta returns compared to large-cap stocks, attracting speculative interest from retail investors [9]. - The lower coverage of small micro-cap stocks by large institutional investors leads to higher mispricing probabilities, providing opportunities for quantitative models to identify undervalued targets [9]. - The current market liquidity favors micro-cap stocks, pushing their prices higher, especially during periods of weak economic data [9]. Group 3: Investor Strategies to Mitigate Risks - As long as micro-cap stocks maintain a strong market position, the likelihood of high exposure in quantitative long products remains significant [10]. - New investors may have concerns, but the current bull market is relatively rare, and any adjustments are expected to manifest as fluctuations rather than sharp declines [10]. - Quantitative long strategies differ from simple micro-cap strategies, focusing on identifying strong stocks and increasing exposure based on market conditions [10].

Quantitative Models and Construction - **Model Name**: Volume Model **Construction Idea**: This model uses trading volume as a key indicator to predict market trends in the short term[12][65] **Construction Process**: The model evaluates the trading volume of broad-based indices to generate buy or sell signals. A higher trading volume relative to historical averages indicates a "bullish" signal, while lower volumes may indicate neutrality or bearishness[12][65] **Evaluation**: The model is effective in capturing short-term market momentum and is widely applicable across broad indices[12][65] - **Model Name**: Low Volatility Model **Construction Idea**: This model focuses on the volatility of indices to assess market stability and predict trends[12][65] **Construction Process**: The model calculates the historical volatility of indices over a defined period. If the volatility is low, the model remains neutral, indicating a stable market environment[12][65] **Evaluation**: The model is useful for identifying periods of market stability but may lack predictive power during high-volatility phases[12][65] - **Model Name**: Institutional Feature Model (Top Trader) **Construction Idea**: This model analyzes institutional trading patterns to predict market movements[12][65] **Construction Process**: The model tracks the trading activity of institutional investors, particularly their buying and selling patterns. A high level of institutional selling generates a "bearish" signal[12][65] **Evaluation**: The model provides insights into institutional sentiment but may be less effective in retail-dominated markets[12][65] - **Model Name**: Momentum Model **Construction Idea**: This model leverages price momentum to predict long-term market trends[14][67] **Construction Process**: The model calculates the rate of price change over a long-term horizon. Positive momentum generates a "bullish" signal, while negative momentum indicates bearishness[14][67] **Evaluation**: The model is effective in identifying long-term trends but may lag during sudden market reversals[14][67] - **Model Name**: A-Share Comprehensive Weapon V3 Model **Construction Idea**: This is a composite model that integrates multiple signals across different time horizons[15][68] **Construction Process**: The model combines short-term, medium-term, and long-term signals from various sub-models (e.g., volume, momentum, institutional features) to generate an overall market outlook[15][68] **Evaluation**: The model balances short-term and long-term perspectives, making it robust for comprehensive market analysis[15][68] - **Model Name**: Hang Seng Turnover-to-Volatility Model **Construction Idea**: This model uses the ratio of turnover to volatility to predict medium-term trends in the Hong Kong market[16][69] **Construction Process**: The model calculates the turnover-to-volatility ratio for the Hang Seng Index. A higher ratio indicates a "bullish" signal, suggesting strong market participation relative to risk[16][69] **Evaluation**: The model is effective in capturing medium-term trends but may be less responsive to short-term fluctuations[16][69] Model Backtesting Results - **Volume Model**: All broad-based indices showed "bullish" signals in the short term[12][65] - **Low Volatility Model**: Neutral signals were observed, indicating stable market conditions[12][65] - **Institutional Feature Model**: Bearish signals were generated due to high institutional selling activity[12][65] - **Momentum Model**: Long-term "bullish" signals were observed, indicating positive price momentum[14][67] - **A-Share Comprehensive Weapon V3 Model**: Overall "bullish" signals were generated, reflecting a positive market outlook[15][68] - **Hang Seng Turnover-to-Volatility Model**: "Bullish" signals were observed, suggesting optimism in the Hong Kong market[16][69]

Quantitative Models and Construction Methods - **Model Name**: GAN_GRU **Model Construction Idea**: The GAN_GRU model combines Generative Adversarial Networks (GAN) for processing volume-price time-series features and Gated Recurrent Unit (GRU) for encoding time-series features to create a stock selection factor[4][13][41] **Model Construction Process**: 1. **GRU Component**: - Input features include 18 volume-price features such as closing price, opening price, turnover, and turnover rate[14][17][19] - Training data consists of the past 400 days of these features, sampled every 5 trading days, forming a 40x18 matrix to predict cumulative returns over the next 20 trading days[18] - Data preprocessing includes outlier removal and normalization at both time-series and cross-sectional levels[18] - Model architecture: Two GRU layers (128, 128) followed by an MLP (256, 64, 64), with the final output being the predicted return (pRet), which serves as the stock selection factor[22] - Training method: Semi-annual rolling training, with training conducted on June 30 and December 31 each year[18] - Optimization: Adam optimizer, learning rate of 1e-4, IC loss function, early stopping after 10 epochs, and a maximum of 50 training epochs[18] 2. **GAN Component**: - GAN consists of a generator (G) and a discriminator (D)[23] - Generator: Uses LSTM to preserve the time-series nature of the input features, transforming random noise into realistic data samples[33][37] - Loss function: $$ L_{G} = -\mathbb{E}_{z\sim P_{z}(z)}[\log(D(G(z)))] $$ where \( z \) represents random noise, \( G(z) \) is the generated data, and \( D(G(z)) \) is the discriminator's output probability[24][25] - Discriminator: Uses CNN to process the two-dimensional volume-price time-series features, distinguishing between real and generated data[33][37] - Loss function: $$ L_{D} = -\mathbb{E}_{x\sim P_{data}(x)}[\log D(x)] - \mathbb{E}_{z\sim P_{z}(z)}[\log(1-D(G(z)))] $$ where \( x \) is real data, \( D(x) \) is the discriminator's output for real data, and \( D(G(z)) \) is the output for generated data[27][29] - Training: Alternating updates of the generator and discriminator parameters until convergence[30] **Model Evaluation**: The GAN_GRU model effectively captures both time-series and cross-sectional features, leveraging the strengths of GAN and GRU for stock selection[4][13][41] --- Model Backtesting Results - **GAN_GRU Model**: - **IC Mean**: 11.36%[41][42] - **ICIR (Non-Annualized)**: 0.88[42] - **Turnover Rate**: 0.83[42] - **Recent IC**: -2.56%[41][42] - **1-Year IC Mean**: 8.94%[41][42] - **Annualized Return**: 38.09%[42] - **Annualized Volatility**: 23.68%[42] - **IR**: 1.61[42] - **Maximum Drawdown**: 27.29%[42] - **Annualized Excess Return**: 23.52%[41][42] --- Quantitative Factors and Construction Methods - **Factor Name**: GAN_GRU Factor **Factor Construction Idea**: Derived from the GAN_GRU model, this factor encodes volume-price time-series features to predict stock returns[4][13][41] **Factor Construction Process**: - The factor is generated using the output of the GAN_GRU model, which combines GAN-based feature generation and GRU-based time-series encoding[4][13][41] - The factor undergoes industry and market capitalization neutralization, as well as standardization, before being used for testing[22] **Factor Evaluation**: The GAN_GRU factor demonstrates strong predictive power across various industries, with consistent outperformance in recent years[4][13][41] --- Factor Backtesting Results - **GAN_GRU Factor**: - **IC Mean**: 11.36%[41][42] - **ICIR (Non-Annualized)**: 0.88[42] - **Turnover Rate**: 0.83[42] - **Recent IC**: -2.56%[41][42] - **1-Year IC Mean**: 8.94%[41][42] - **Annualized Return**: 38.09%[42] - **Annualized Volatility**: 23.68%[42] - **IR**: 1.61[42] - **Maximum Drawdown**: 27.29%[42] - **Annualized Excess Return**: 23.52%[41][42]

金融工程 金工周报 2025 年 08 月 24 日 【金工周报】（20250818-20250822） 大部分指数依旧看多，后市或乐观向上 本周回顾 本周市场普遍上涨，上证指数单周上涨 3.49%，创业板指单周上涨 5.85%。 A 股模型： 短期：成交量模型大部分宽基看多。低波动率模型中性。特征龙虎榜机构模型 看空。特征成交量模型看多。智能算法沪深 300 模型看多，智能算法中证 500 模型看多。 证 券 研 究 报 告 中期：涨跌停模型看多。月历效应模型中性。 长期：长期动量模型看多。 综合：A 股综合兵器 V3 模型看多。A 股综合国证 2000 模型看多。 港股模型： 中期：成交额倒波幅模型看多。 本周行业指数普遍上涨，涨幅前五的行业为：通信、电子、计算机、传媒、综 合。从资金流向角度来说，除通信、消费者服务、综合外所有行业主力资金净 流出，其中机械、医药、计算机、电力设备及新能源、基础化工主力资金净流 出居前。 本周股票型基金总仓位为 98.71%，相较于上周减少了 40 个 bps，混合型基金 总仓位 95.36%，相较于上周增加了 264 个 bps。 本周电力设备及新能源与通信获得最大机构 ...

在发展方向上，团队聚焦量化选股与资产配置两大领域：量化选股方面，采用人工 + 智能模式，打造 稳定的增强工具，涵盖传统多因子策略、基本面量化策略、AI 策略等，并持续优化因子评价、选择、 迭代机制；资产配置方面，打造高夏普、工程化解决方案，从自上而下视角出发，以强逻辑基本面框架 为支撑，涉及股票风格配置与行业配置、债券量化配置、大类资产配置等。 8月18日，南方基金旗下南方中证A500指数增强型证券投资基金(基金简称：南方中证A500指数增强； 基金代码：A类024375，C类024376)正式发售。该基金锚定新一代大盘宽基指数——中证 A500 指数， 依托南方基金数量化投资团队的积淀与量化多策略赋能，在跟踪指数 Beta 收益的基础上力争超额收 益，为投资者提供布局 A 股核心资产的优质工具。 凝聚团队力量 打造α增强引擎 据了解，南方中证 A500 指数增强的第一大亮点，在于其背后团队的强力支撑。据了解，南方基金数量 化投资团队由 13 位成员组成的专业团队，成员涵盖数学、金融工程、信息技术等复合背景，平均从业 时间超 8 年，投资团队平均从业时间逾 10 年，具备深厚的专业功底和丰富的实战经验。 近年 ...

Group 1 - The core viewpoint is that the private equity market is experiencing a significant recovery, with over 90% of large private equity firms achieving positive returns this year, driven by structural market opportunities and active trading [2][3] - As of the end of July, the average return for large private equity firms was reported at 16.6%, with 54 out of 55 firms showing positive returns, indicating a strong performance in the sector [2] - The number of large private equity firms has increased to 90, reflecting the expansion of the industry amid favorable market conditions [1][2] Group 2 - The issuance market for private equity has notably improved, with a total of 1,298 private equity securities investment funds registered in July, marking an 18% increase from the previous month [3] - Large private equity firms dominated the new fund registrations in July, with significant numbers of new funds being launched, particularly in index-enhanced strategies [3] - Investor sentiment has improved, with institutional investors increasing their participation and shifting their preferences towards long-biased strategies, while individual investors are also showing signs of renewed interest [3] Group 3 - Large private equity firms are maintaining aggressive positions and actively adjusting their portfolios to capitalize on structural opportunities in the market [4][5] - The current investment focus includes sectors such as technology, innovative pharmaceuticals, non-bank financials, and cyclical stocks, with a high portfolio allocation of over 80% [4] - There is an expectation of profit-taking in popular sectors due to recent gains, particularly during the busy earnings reporting period in August, leading to potential adjustments in investment strategies [5]

Quantitative Models and Construction - **Model Name**: Volume Model **Construction Idea**: This model evaluates market trends based on trading volume changes over time [12][72] **Construction Process**: The model analyzes the trading volume of broad-based indices to determine short-term market sentiment. It transitions between "bullish," "neutral," and "bearish" signals based on volume dynamics [12][72] **Evaluation**: The model is effective in capturing short-term market sentiment but may require integration with other indicators for comprehensive analysis [12][72] - **Model Name**: Low Volatility Model **Construction Idea**: This model assesses market conditions by analyzing the volatility of indices [12][72] **Construction Process**: The model calculates the historical volatility of indices and assigns a "neutral" signal when volatility remains within a predefined range [12][72] **Evaluation**: The model provides a stable perspective on market conditions but may lag in highly volatile environments [12][72] - **Model Name**: Intelligent Algorithm Model (CSI 300 and CSI 500) **Construction Idea**: This model uses machine learning algorithms to predict market trends for specific indices [12][72] **Construction Process**: The model applies advanced algorithms to historical price and volume data, generating "bullish" signals for the CSI 300 and CSI 500 indices [12][72] **Evaluation**: The model demonstrates strong predictive capabilities for these indices, particularly in short-term scenarios [12][72] - **Model Name**: Limit-Up/Limit-Down Model **Construction Idea**: This model evaluates market sentiment based on the frequency of limit-up and limit-down events [13][73] **Construction Process**: The model tracks the number of stocks hitting daily price limits and assigns a "neutral" signal when no significant trend is observed [13][73] **Evaluation**: The model is useful for identifying extreme market conditions but may not capture subtle trends [13][73] - **Model Name**: Long-Term Momentum Model **Construction Idea**: This model identifies long-term trends by analyzing momentum indicators [14][74] **Construction Process**: The model calculates momentum metrics for indices like the SSE 50, which recently transitioned to a "bullish" signal [14][74] **Evaluation**: The model is effective for long-term trend analysis but may miss short-term fluctuations [14][74] - **Model Name**: A-Share Comprehensive Weapon V3 Model **Construction Idea**: This composite model integrates multiple signals to provide an overall market outlook [15][75] **Construction Process**: The model aggregates signals from various short-term, medium-term, and long-term models, currently indicating a "bearish" outlook [15][75] **Evaluation**: The model offers a holistic view but may dilute the impact of individual signals [15][75] - **Model Name**: HK Stock Turnover-to-Volatility Model **Construction Idea**: This model evaluates the Hong Kong market by analyzing turnover relative to volatility [16][76] **Construction Process**: The model calculates the ratio of turnover to volatility, currently signaling a "bullish" outlook for the Hang Seng Index [16][76] **Evaluation**: The model is effective for medium-term analysis but may require additional factors for short-term predictions [16][76] Model Backtesting Results - **Volume Model**: Short-term signal transitioned to "neutral" for most broad-based indices [12][72] - **Low Volatility Model**: Maintains a "neutral" signal [12][72] - **Intelligent Algorithm Model**: "Bullish" signals for CSI 300 and CSI 500 indices [12][72] - **Limit-Up/Limit-Down Model**: "Neutral" signal for medium-term analysis [13][73] - **Long-Term Momentum Model**: SSE 50 transitioned to "bullish" [14][74] - **A-Share Comprehensive Weapon V3 Model**: Overall "bearish" signal [15][75] - **HK Stock Turnover-to-Volatility Model**: "Bullish" signal for the Hang Seng Index [16][76]