Market Performance - The Sci-Tech 50 Index decreased by 4.03% over the last five trading days, while the ChiNext Index increased by 1.26%. The large-cap value index fell by 1.44%, and the large-cap growth index dropped by 0.48%. The Shanghai 50 Index declined by 2.47%, and the small-cap index represented by the CSI 2000 fell by 5.45%. The communication and banking sectors performed well, while basic chemicals and non-ferrous metals lagged behind [1]. Valuation Levels - As of March 20, 2026, the static PE of the CSI All Share Index is at 2.63%, with a two-standard deviation boundary at 4.62%. The CSI All Share Index's TTM PE is at the 82nd percentile, while the Shanghai 50 and CSI 300 are at 70% and 73%, respectively. The ChiNext Index is close to 63%, and the CSI 500 and CSI 1000 are both at 65%. The ChiNext Index's valuation style is relatively at the historical median level [2]. Thematic Investment Strategy - The latest thematic investment strategy focuses on low volatility dividends, banking, and real estate sectors. Specific indices include the CSI Low Volatility Dividend Index, CSI Banking Index, CSI 800 Banking Index, CSI Mainland Real Estate Theme Index, and CSI 800 Real Estate Index [2][3]. AI and Machine Learning Application - A convolutional neural network (CNN) is utilized to model price and volume data, mapping learned features to industry thematic sectors. This approach is based on research reports regarding AI recognition and classification of stock price trends [9]. Market Sentiment - The proportion of stocks above the 200-day long-term moving average is tracked, indicating market sentiment and potential trends [10]. ETF Scale Changes - The report includes observations on the changes in the scale of mainstream ETFs, reflecting shifts in investor preferences and market dynamics [11]. Risk Preference Tracking - The report monitors the risk preferences between equity and bond assets, providing insights into investor behavior and market conditions [12].

Market Performance - The Sci-Tech 50 Index decreased by 2.88% over the last five trading days, while the ChiNext Index increased by 2.51%. The large-cap value index rose by 0.48%, and the large-cap growth index increased by 1.38%. The Shanghai 50 Index fell by 1.20%, and the small-cap index represented by the CSI 2000 declined by 1.13%. Coal and electric equipment sectors performed well, while defense, military, oil, and petrochemical sectors lagged behind [1]. Risk Premium and Valuation Levels - As of March 13, 2026, the risk premium, measured as the inverse of the static PE of the CSI All Share Index minus the yield of 10-year government bonds, stood at 2.48%. The two-standard deviation boundary is 4.63% [1]. - The valuation level of the CSI All Share Index's TTM PE is at the 83rd percentile. The Shanghai 50 and CSI 300 indices are at 71% and 75%, respectively, while the ChiNext Index is close to 62%. The CSI 500 and CSI 1000 indices are at 68% and 67%, respectively, indicating that the ChiNext Index's valuation is relatively at the historical median level [1]. ETF Fund Flow - In the last five trading days, ETF funds experienced an outflow of 12.3 billion yuan, while margin trading increased by approximately 12 billion yuan. The average daily trading volume across both markets was 24,792 billion yuan [2]. Industry Themes and Indexes - The latest thematic allocation focuses on sectors such as electricity, power grid, and public utilities, including specific indices like the National Green Power Index, CSI Green Power Index, CSI All Share Power Public Utilities Index, CSI All Share Public Utilities Index, and CSI Power Grid Equipment Theme Index [2][3][12]. Market Sentiment and Risk Preference - The report includes observations on market sentiment based on the proportion of stocks above the 200-day long-term moving average, as well as tracking risk preferences between equity and bond assets [13][14]. Financing Balance - The report provides insights into the financing balance, indicating trends in margin trading and overall market leverage [16]. Individual Stock Performance - There is a statistical distribution of individual stock performance year-to-date based on return intervals, highlighting the performance of various stocks within the market [18]. Oversold Indices - The report notes instances of indices being oversold, which may present potential investment opportunities [20].

Market Performance - The Sci-Tech 50 Index decreased by 4.95% over the last five trading days, while the ChiNext Index fell by 2.45%. In contrast, the large-cap value stocks rose by 1.17%, and large-cap growth stocks declined by 1.10%. The Shanghai Stock Exchange 50 Index dropped by 1.54%, whereas the small-cap index represented by the CSI 2000 increased by 3.53%. The oil and coal sectors performed well, while media and non-ferrous metals lagged behind [1]. Valuation Levels - As of March 6, 2026, the static PE of the CSI All Share Index indicates an earnings yield (EP) of 2.47% when compared to the 10-year government bond yield. The two-standard deviation boundary is at 4.64%. The CSI All Share Index's PE TTM is at the 84th percentile, with the SSE 50 and CSI 300 at 72% and 75%, respectively. The ChiNext Index is close to 58%, while the CSI 500 and CSI 1000 are at 69% and 68%, respectively. The ChiNext Index's valuation is relatively at the historical median level [1]. ETF Fund Flows - In the last five trading days, ETF inflows amounted to 5 billion yuan, while the financing balance decreased by approximately 16 billion yuan. The average daily trading volume across the two markets was 26,212 billion yuan [2]. Industry Themes - The latest thematic allocation focuses on sectors such as electricity, power grids, and public utilities. This includes specific indices like the National Green Power Index, CSI Green Power Index, CSI All Share Power Utility Index, CSI All Share Utility Index, and CSI Power Grid Equipment Theme Index [2][3]. Long-term Market Sentiment - The proportion of stocks above the 200-day moving average is being tracked to gauge market sentiment [13]. Risk Preference Tracking - The risk preference between equity and bond assets is being monitored, reflecting investor sentiment towards riskier assets [14]. Financing Balance - The financing balance data indicates trends in investor leverage and market participation [16]. Individual Stock Performance - Statistics on individual stock performance year-to-date based on return ranges are being compiled to assess market dynamics [19]. Oversold Indices - Analysis of indices that are currently oversold is being conducted to identify potential buying opportunities [20].

Market Performance - The Sci-Tech 50 Index increased by 0.47% over the last five trading days, while the ChiNext Index decreased by 0.53%. The large-cap value index fell by 1.34%, and the large-cap growth index dropped by 0.93%. The Shanghai Stock Exchange 50 Index declined by 1.31%, whereas the small-cap index represented by the CSI 2000 rose by 3.08%. The steel and environmental sectors performed well, while media and non-bank financial sectors lagged behind [1]. Risk Premium and Valuation Levels - As of February 27, 2026, the risk premium, measured as the inverse of the static PE of the CSI All Share Index minus the yield of ten-year government bonds, stands at 2.43%. The two standard deviation boundary is 4.65% [1]. - The valuation levels indicate that the CSI All Share Index's PETTM is at the 84th percentile. The Shanghai 50 and CSI 300 are at 72% and 74%, respectively, while the ChiNext Index is close to 61%. The CSI 500 and CSI 1000 are at 70% and 69%, respectively, indicating that the ChiNext Index's valuation is relatively at the historical median level [1]. ETF Fund Flow - In the last five trading days, there was an outflow of 39.3 billion yuan from ETFs, while the margin trading balance increased by approximately 22.2 billion yuan. The average daily trading volume across the two markets was 23,348 billion yuan [2]. Industry Themes and Indices - The latest thematic allocation includes industries such as shipbuilding, electric power, steel, and robotics, specifically represented by indices like the CSI Selected Shipbuilding Industry Index, CSI Electric Power Equipment Theme Index, CSI Steel Index, and CSI Robotics Index [2][3].

- The report discusses the use of convolutional neural networks (CNNs) to model price-volume data and future price trends, transforming these learned features into industry theme indices such as the CSI Smart Shipbuilding Industry Index, CSI Power Grid Equipment Theme Index, CSI Steel Index, and CSI Robotics Index[81][82][87] - The CNN model constructs standardized charts of price-volume data within specific time windows for individual stocks, which are then used to train the model to identify patterns and predict future price movements[81][82] - The CNN model's thematic allocation currently focuses on sectors like shipbuilding, power grids, steel, and robotics, as reflected in the indices mentioned above[81][82][87]

Summary of Key Points Core Insights - The report focuses on an ETF rotation strategy based on convolutional neural networks (CNN), utilizing the Wind industry thematic indices as the underlying assets. The strategy aims to measure the effectiveness of rotation based on these indices [4][9]. - The Wind industry thematic indices include over 1,000 indices across various categories, providing a broader selection compared to traditional ETFs, which have around 419 tracked indices as of January 2026 [4][45]. Section Summaries 1. Background Introduction - The report highlights the increasing acceptance of index-based investment strategies, particularly ETFs, which are favored for their transparency, low fees, and ease of trading. The strategy discussed has shown significant excess returns compared to the Wind mixed equity fund index since its implementation [9]. 2. Convolutional Neural Network Factor Logic - The methodology involves creating standardized price-volume data charts for stocks, which are then used to train a CNN model to predict future stock price movements. The model processes a large dataset of 115GB, significantly larger than traditional sequential data [13][19]. 3. Wind Industry Thematic Index Information - The Wind industry thematic indices are categorized into four subtypes: industry, theme, popular concepts, and thematic indices, with a total of 1,046 indices as of January 2026. This extensive categorization allows for a more nuanced investment approach compared to ETFs [27][45]. 4. Empirical Analysis - The empirical analysis indicates that the CNN-based rotation strategy achieved an average annualized return of approximately 30.7% since 2020, outperforming the Wind mixed equity fund index by about 21.7%. The strategy's IC (Information Coefficient) average is 3.7%, with a win rate of 59% [54][56]. - The analysis also shows that the strategy's performance is more stable across years compared to ETF rotation strategies, although the overall return in 2025 was lower than that of the ETF rotation [55][63]. 5. Parameter Adjustment Impact Measurement - The report examines the impact of various parameters on the strategy's performance, including the number of holdings (3, 5, or 10 indices), turnover frequency (weekly, bi-weekly, or monthly), and the price at which trades are executed. The findings suggest that a weekly turnover strategy yields higher returns [55][61]. 6. Conclusion - The report concludes that the CNN-based ETF rotation strategy, leveraging the diverse Wind industry thematic indices, presents a promising investment opportunity with significant potential for excess returns compared to traditional ETF strategies [4][9].

Market Performance - The Sci-Tech 50 Index decreased by 2.85% over the last five trading days, while the ChiNext Index fell by 0.09%. In contrast, the large-cap value index rose by 1.87%, and the large-cap growth index increased by 0.68%. The Shanghai 50 Index gained 1.13%, whereas the small-cap index represented by the CSI 2000 dropped by 2.76%. The telecommunications and oil & petrochemical sectors performed well, while the defense, military, and power equipment sectors lagged behind [1]. Valuation Levels - As of January 30, 2026, the static PE of the CSI All Share Index is at 2.49%, with a two-standard deviation boundary of 4.67%. The valuation levels indicate that the CSI All Share Index's TTM PE is at the 84th percentile, while the Shanghai 50 and CSI 300 are at 74% and 75%, respectively. The ChiNext Index is close to 62%, and the CSI 500 and CSI 1000 are at 69% and 67%, respectively, suggesting that the ChiNext Index's valuation is relatively at the historical median level [2]. Fund Flow and Trading Activity - In the last five trading days, ETF funds experienced an outflow of 316.7 billion yuan, while the margin trading balance increased by approximately 14.7 billion yuan. The average daily trading volume across the two markets was 30.348 billion yuan [2]. Thematic Investment Insights - The latest thematic investment focus includes sectors such as petrochemicals, chemicals, and non-ferrous metals. Specific indices highlighted are the CSI Petrochemical Industry Index, CSI Sub-Industry Chemical Theme Index, CSI Oil and Gas Index, and CSI Non-Ferrous Index [2][3]. AI and Machine Learning Applications - The report discusses the application of convolutional neural networks (CNN) to model price and volume data, aiming to identify future price movements and map learned features to industry themes. This approach is based on research reports related to AI recognition and classification of stock price trends [2][11].

- The report introduces a quantitative model based on Convolutional Neural Networks (CNNs) to analyze price-volume data and predict future prices. The model standardizes price-volume data into graphical representations and maps learned features to industry theme indices, such as the CSI Petrochemical Industry Index, CSI Subdivision Chemical Industry Theme Index, CSI Machine Tool Index, CSI Semiconductor Material Equipment Theme Index, and CSI Nonferrous Metals Index[78][80][81] - The construction process of the CNN model involves transforming individual stock price-volume data within a specific window into standardized graphical charts. These charts are then input into the CNN for feature extraction and prediction modeling. The learned features are subsequently applied to identify and allocate industry themes[78][80] - The evaluation of the CNN model highlights its ability to capture complex patterns in price-volume data and effectively map these patterns to industry themes. This approach provides a novel perspective for quantitative investment strategies[78][81] - Backtesting results indicate that the CNN model's latest configuration suggests a focus on themes such as petrochemicals, chemicals, machine tools, semiconductors, and nonferrous metals. Specific indices include the CSI Petrochemical Industry Index, CSI Subdivision Chemical Industry Theme Index, CSI Machine Tool Index, CSI Semiconductor Material Equipment Theme Index, and CSI Nonferrous Metals Index[80][81]

Market Performance - The Sci-Tech 50 Index increased by 2.62% over the last five trading days, while the ChiNext Index decreased by 0.34%. The large-cap value index fell by 1.64%, and the large-cap growth index dropped by 1.34%. The Shanghai 50 Index declined by 1.54%, whereas the small-cap index represented by the CSI 2000 rose by 3.33%. The building materials and oil & petrochemical sectors performed well, while banks and telecommunications lagged behind [1]. Risk Premium and Valuation Levels - As of January 23, 2026, the static PE of the CSI All Share Index minus the yield of 10-year government bonds indicates a risk premium of 2.46%, with a two-standard deviation boundary at 4.68%. The valuation levels show that the CSI All Share Index's PETTM is at the 84th percentile, with the Shanghai 50 and CSI 300 at 72% and 73%, respectively. The ChiNext Index is close to 63%, while the CSI 500 and CSI 1000 are at 70% and 68%, respectively. The ChiNext Index's valuation is relatively at the historical median level [1]. ETF Fund Flow - In the last five trading days, ETF funds experienced an outflow of 326.5 billion yuan, while the financing balance increased by approximately 6.5 billion yuan. The average daily trading volume across the two markets was 27.727 billion yuan [2]. Industry Themes and Indices - The latest thematic allocation includes sectors such as petrochemicals, chemicals, machine tools, semiconductors, and non-ferrous metals. Specific indices mentioned are the CSI Petrochemical Industry Index, CSI Sub-segment Chemical Industry Theme Index, CSI Machine Tool Index, CSI Semiconductor Materials and Equipment Theme Index, and the National Non-ferrous Metals Index [2][3]. Long-term Market Sentiment - The report includes observations on the proportion of stocks above the 200-day long-term moving average, indicating market sentiment trends [13]. Financing Balance - The report tracks the financing balance, which is a critical indicator of market liquidity and investor sentiment [16]. Individual Stock Performance - A statistical distribution of individual stocks based on their return ranges since the beginning of the year is provided, highlighting performance variations across different stocks [18]. Oversold Indices - The report notes instances of indices being oversold, which may present potential buying opportunities [20].

Market Performance - The Sci-Tech 50 Index increased by 2.58% over the last five trading days, while the ChiNext Index rose by 1.00%. In contrast, the large-cap value index fell by 2.81%, and the large-cap growth index decreased by 0.03%. The Shanghai Stock Exchange 50 Index dropped by 1.74%, whereas the small-cap index represented by the CSI 2000 gained 1.31%. The computer and electronics sectors performed well, while the defense, military, and real estate sectors lagged behind [1]. Risk Premium and Valuation Levels - As of January 16, 2026, the risk premium, measured as the inverse of the static PE of the CSI All Share Index minus the yield of ten-year government bonds, stands at 2.51%. The two standard deviation boundary is 4.68%. The valuation levels indicate that the CSI All Share Index's PETTM is at the 83rd percentile, with the Shanghai 50 and CSI 300 at 74% and 75%, respectively. The ChiNext Index is close to 63%, while the CSI 500 and CSI 1000 are at 69% and 67%, respectively. The ChiNext Index's valuation is relatively at the historical median level [1]. Fund Flow and Trading Volume - In the last five trading days, ETF funds experienced an outflow of 132.3 billion yuan, while the margin trading balance increased by approximately 97.3 billion yuan. The average daily trading volume across the two markets was 34.251 billion yuan [3]. Thematic Indexes - The latest thematic allocations include the Satellite Industry Index, Nonferrous Metals Index, Biotechnology Index, and Computer Index, among others. Specific indices mentioned are the CSI Satellite Industry Index, CSI Industrial Nonferrous Metals Theme Index, CSI Biotechnology Theme Index, CSI Big Data Industry Index, and CSI Computer Theme Index [2][4]. Market Sentiment and Risk Preference - The report includes observations on market sentiment based on the proportion of stocks above the 200-day moving average and tracks the risk preferences between equity and bond assets [13][14]. Financing Balance - The financing balance data is also included, indicating trends in margin trading activity [16]. Individual Stock Performance - Statistics on individual stock performance year-to-date based on return intervals are provided, highlighting the distribution of returns across different stocks [18]. Oversold Indices - The report notes instances of indices being oversold, which may indicate potential investment opportunities [20].