| | | [Table_Report] 相关报告 服务消费景气提升，科技硬件延续涨价 [Table_Authors] 方奕(分析师) 中观景气 12 月第 1 期 本报告导读： 中观景气延续分化的增长格局，新兴科技景气仍强，高性能存储价格延续快速上涨， 游戏供给偏宽松；服务消费景气明显提升，地产周期和耐用品需求仍承压。 投资要点： | | 021-38031658 | | --- | --- | | | fangyi2@gtht.com | | 登记编号 | S0880520120005 | | | 陶前陈(研究助理) | | | 0755-23976164 | | | taoqianchen@gtht.com | | 登记编号 | S0880125070014 | | | 张逸飞(分析师) | 资产概览：银价铜价创历史新高 2025.12.01 外资与融资资金重回流入 2025.12.01 消费景气线索增多，科技制造延续增长 2025.11.26 ETF 流入明显，融资资金与外资有所流出 2025.11.24 资产概览：美元兑日元逼近 160 关口 2025.11.24 策 略 研 究 告 请务必阅读正文 ...

Report Industry Investment Rating No relevant content provided. Core Viewpoints of the Report - The timing model constructed in a "scenario-based combat" approach in 2025 performed averagely due to factors such as unstable effectiveness, multicollinearity, and high volatility. A new volume-price factor timing model based on a grouping algorithm was reconstructed, focusing on optimizing three major issues: unstable effectiveness, large signal fluctuations, and insufficient reliability of single signals [4][7]. - By using double standards of grouped IC and thresholds in factor screening, factors that can stably play a predictive role in both high and low value intervals were selected, ensuring the effectiveness of model information from the source [4][10]. - Through the strategies of "random grouping + independent training" and "rolling smoothing + multi - group voting", noise was filtered, effective information was aggregated, and accurate and robust timing signals were generated [4]. - The back - test results showed that the model significantly outperformed the benchmark, especially the 1 - day signal, which demonstrated strong stable timing ability both within and outside the sample [4]. Summary According to the Directory 1. Factor Screening: Double Standards Determine Stability and Effectiveness - **Factor Reserve**: 87 factors covering intraday patterns, price fluctuations, trading volume statistics, trends, momentum, overbought and oversold conditions were constructed around the volume - price characteristics of Treasury bond futures, comprehensively covering core dimensions for timing [11]. - **Screening Standard**: An annual rolling back - test framework was adopted, with a 3 - year data window each year. Factors were sorted and divided into 5 groups. Two conditions were set: at least 4 groups of IC should maintain the same direction, and the average absolute value of the same - direction IC should be no less than 0.05. This mechanism could identify factors with stable cross - sectional prediction ability and adapt to market changes [13][14]. 2. Model Building and Signal Generation: Group Voting Based on Random Grouping and Cross - Validation - **Model Building**: A deep - learning architecture of bidirectional multi - layer GRU + attention mechanism was used, focusing on short - term timing requirements for T + 1 and T + 5 day price movements. Techniques such as Dropout and layer normalization were introduced to avoid overfitting, and the "rolling window" training mode was adopted [16]. - **Random Grouping and Signal Generation** - **Random Grouping of Factors and Parallel Training**: After selecting effective factors based on grouped IC values each year, they were randomly divided into multiple groups. Each group of factors was used to train an independent GRU sub - model, generating diverse prediction results [20]. - **Signal Generation**: The final timing signal was generated through "rolling smoothing + multi - group voting". Rolling smoothing was used to filter noise according to different prediction periods, and multi - group voting was used to confirm the signal direction, reducing the influence of single - sub - model prediction errors. A comprehensive signal was also generated by integrating 1 - day and 5 - day prediction results [21][22]. 3. Strategy Back - test Design and Parameter Selection: Balancing Robustness and Adaptability - **Strategy Back - test Design**: A long - short full - position trading mode was set. When the final signal was 1, a full - position long was taken; when it was - 1, a full - position short was taken; when it was 0, the position remained unchanged. Transaction costs and slippage were ignored, and the closing price of the 10 - year Treasury bond futures main contract was used as the benchmark [25]. - **Parameter Selection**: Three major dimensions of parameters were focused on: training window, number of groups, and number of factors in each group. Different parameter combinations were tested from 2019 to September 2025, and the optimal parameter combinations were selected based on the back - test results, emphasizing overall return - capturing ability and stable advantages over the benchmark in most years [26]. - **Back - test Performance**: The 1 - day signal performed outstandingly both within and outside the sample. Within the sample, the 1 - day signal had an average annualized return of 3.61% and a Sharpe ratio of 1.12; outside the sample, from the National Day holiday in 2025 to the present, the 1 - day signal had a cumulative return of 0.99%, a Sharpe ratio of 5.98, and a maximum drawdown of only 0.13%. The 5 - day signal's adaptability decreased outside the sample [30][34].

Investment Rating - The report assigns an "Accumulate" rating for the industry [6] Core Insights - The launch of the AI mobile assistant by Doubao marks a significant step towards the era of AI smartphones, enhancing the hardware ecosystem [2][6] - Doubao's AI assistant integrates deeply with the operating system, allowing for complex command execution and seamless user interaction across applications [6] - The report highlights the potential for Doubao to provide AI capabilities to mid-range smartphone manufacturers, enhancing their competitiveness in the AI era [6] Summary by Sections Industry Overview - The report discusses the rapid emergence of AI smartphones and the growing ecosystem of edge hardware, driven by innovations from major manufacturers [2][6] Investment Recommendations - Recommended stocks include Luxshare Precision, GoerTek, and Pengding Holdings, with related stocks being Transsion Holdings, Lens Technology, and Haopeng Technology [6][7] Technology Insights - Doubao's AI assistant can understand and interact with the content displayed on the user's screen, facilitating complex tasks such as price comparisons and travel planning [6] - The assistant operates without interrupting ongoing app processes, enhancing user experience [6] Market Potential - The report emphasizes the opportunity for Doubao to collaborate with smartphone manufacturers, creating a business model that combines AI capabilities with hardware [6] - The edge hardware ecosystem is expected to grow rapidly, with Doubao's technology enabling usability in devices like AI glasses and TWS earbuds [6]

Investment Rating - The report does not explicitly state an investment rating for the industry [2] Core Insights - The Shenzhen-Singapore cross-border data verification platform was launched on November 25, 2025, utilizing the FISCO BCOS blockchain technology to enhance data security and compliance in cross-border transactions [7][9] - The platform aims to facilitate two-way credit report verification between Guangdong and Singapore, leveraging the successful practices established in the Greater Bay Area [9][17] - FISCO BCOS serves as the underlying technology, providing a secure and efficient environment for digital asset applications and ensuring data integrity through hash verification [10][13] - The collaboration with Experian Singapore enhances the platform's credibility and acceptance in the international credit system, potentially becoming a key point for broader cross-border credit cooperation between the Greater Bay Area and ASEAN countries [16][17] Summary by Sections 1. Launch of the Shenzhen-Singapore Cross-Border Data Verification Platform - The platform operates on a model of "user self-transmission, platform hash verification," effectively separating original data from verification information to address compliance and security challenges in cross-border data flow [7][9] - It connects over 40 institutions and has implemented more than 15 use cases, covering the 9+2 city clusters in the Greater Bay Area [8][9] 2. FISCO BCOS as the Underlying Technology - FISCO BCOS is a domestic open-source blockchain platform designed for financial applications, supporting over 400 digital transformation projects across various sectors [10][11] - The platform's high performance and security features, including support for multiple consensus algorithms and privacy protection mechanisms, make it suitable for enterprise-level applications [11][12] 3. Experian Singapore's Role in ASEAN Cooperation - Experian Singapore, a key player in the local credit information infrastructure, enhances the platform's capability for cross-border credit verification [16][17] - The partnership is expected to improve the acceptance of verification results within Singapore's regulatory framework, facilitating broader credit cooperation in the region [17]

守正待变：数据真空下中久期高评级策略 本报告导读： 在不确定性窗口期以中久期高评级策略平衡收益获取与风险控制。 投资要点： 债券研究 /[Table_Date] 2025.12.03 | [Table_Authors] | 王一凡(分析师) | | --- | --- | | | 021-38031722 | | | wangyifan4@gtht.com | | 登记编号 | S0880524120001 | | | 唐元懋(分析师) | | | 0755-23976753 | | | tangyuanmao@gtht.com | | 登记编号 | S0880524040002 | [Table_Report] 相关报告 从超跌到反弹，技术指标指示了什么 20251130 2025.12.01 年内新增地方债额度基本发行完毕 2025.12.01 11 月最后一周，大行融出回升但未超季节性 2025.12.01 信用债利差走阔：波动还是趋势 2025.11.30 商业不动产 REITs提上日程 2025.11.30 证 券 研 究 报 告 请务必阅读正文之后的免责条款部分 债 券 研 究 市 场 策 略 周 ...

Investment Rating - The report maintains a rating of "Accumulate" for the company, with a target price of 16.50 yuan [6][13]. Core Insights - The controlling shareholder, China Energy Conservation, plans to increase its stake in the company by 365 to 730 million yuan, reflecting confidence in the company's future development prospects [2][13]. - The company achieved a revenue of 2.826 billion yuan in the first three quarters of 2025, representing a year-on-year growth of 2.31% [13]. - The company is actively expanding into the new materials sector, with significant projects underway, including the second phase of the Wanrun Industrial Park, which will produce semiconductor-related materials [13]. Financial Summary - Total revenue is projected to decline from 4.305 billion yuan in 2023 to 3.693 billion yuan in 2024, before recovering to 4.857 billion yuan by 2027, with a CAGR of 13.6% from 2025 to 2027 [4][14]. - Net profit attributable to the parent company is expected to decrease significantly from 763 million yuan in 2023 to 246 million yuan in 2024, before increasing to 577 million yuan by 2027 [4][14]. - Earnings per share (EPS) are forecasted to rise from 0.27 yuan in 2024 to 0.63 yuan in 2027 [4][14]. Market Data - The company's market capitalization is approximately 13.069 billion yuan, with a 52-week stock price range of 9.79 to 14.39 yuan [7][8]. - The company has a net asset value per share of 7.80 yuan, with a price-to-book ratio of 1.8 [8][14]. Industry Context - The company operates within the basic chemicals and raw materials sector, focusing on new materials [5].

Macro Research - The outlook for 2026 suggests a potential diminishing multiplier effect from durable goods subsidies, with price recovery largely dependent on effective fiscal input, particularly towards consumption and services [2][4] - In a neutral scenario, both CPI and PPI are expected to see a moderate upward shift [2][4] Retail and Service Research - The company Emei Mountain A is expected to benefit from the renovation of the Golden Summit cableway and the construction of Leshan Airport, enhancing the scenic area's capacity and visitor experience, which will drive traffic growth and profit recovery [2][6] - The renovation of the Golden Summit cableway will increase its capacity from approximately 1,200 people per hour to 3,200 people per hour, representing a 167% increase, which will alleviate peak waiting times and enhance visitor spending [8] - Leshan Airport is projected to facilitate tourist growth, with a designed annual throughput of 2.6 million passengers, expected to connect with existing highway networks to create a direct transport corridor to major scenic spots [8] Company Coverage - Emei Mountain A is rated as "Buy" with a target price of 20.65 CNY, compared to the current price of 14.46 CNY, reflecting a market capitalization of 761.9 million [7] - The company is expected to achieve net profits of 251 million CNY, 308 million CNY, and 346 million CNY for the years 2025-2027, with corresponding EPS of 0.48 CNY, 0.59 CNY, and 0.66 CNY [7] - The company has established a diversified business ecosystem through a full industry chain layout, including ticket sales, cableway operations, hotels, and various derivative businesses [7] Market Potential - Emei Mountain's visitor numbers were 4.67 million in 2024, significantly lower than other comparable sites, indicating substantial room for market growth through improved transportation and product upgrades [9] - The company has demonstrated strong unit visitor value conversion capabilities, with total tourism revenue of 1.013 billion CNY in 2024, higher than that of Jiuhua Mountain [9]

- The report introduces two key tracking indicators: "Large Order Transaction Amount Proportion" and "Net Active Buy Amount Proportion" [7] - The "Large Order Transaction Amount Proportion" is designed to capture the buying behavior of large funds. It is calculated by identifying large orders from transaction data and computing the proportion of large buy orders' transaction amounts relative to the total daily transaction amount [7] - The "Net Active Buy Amount Proportion" reflects investors' active buying behavior. It is derived by distinguishing active buy and sell transactions from transaction data, calculating the net active buy amount (active buy amount minus active sell amount), and expressing it as a proportion of the total daily transaction amount [7] Factor Backtesting Results - For individual stocks, the top 10 stocks with the highest 5-day average "Large Order Transaction Amount Proportion" include Xinhua Du (90.6%, 99.2% percentile), Beichen Industrial (89.1%, 98.8% percentile), and Zhongyou Engineering (88.8%, 100.0% percentile) [9] - For individual stocks, the top 10 stocks with the highest 5-day average "Net Active Buy Amount Proportion" include Senying Windows (22.3%, 100.0% percentile), Huitong Group (20.0%, 100.0% percentile), and Yuandong Biotech (19.6%, 100.0% percentile) [10] - For broad-based indices, the 5-day average "Large Order Transaction Amount Proportion" ranges from 71.7% (Shanghai 50 Index) to 74.3% (China Securities 500 Index), while the "Net Active Buy Amount Proportion" ranges from -5.2% (Shanghai 50 Index) to 1.9% (China Securities 500 Index) [12] - For industries, the 5-day average "Large Order Transaction Amount Proportion" is highest in the banking sector (80.6%, 86.5% percentile) and lowest in the electronics sector (70.8%, 16.4% percentile). The "Net Active Buy Amount Proportion" is highest in the steel sector (7.9%, 75.8% percentile) and lowest in the banking sector (-14.6%, 3.3% percentile) [13] - For ETFs, the top 10 ETFs with the highest 5-day average "Large Order Transaction Amount Proportion" include Guotai CSI A500 ETF (92.3%, 36.1% percentile) and Guotai SSE 10-Year Treasury Bond ETF (90.7%, 89.3% percentile) [15] - For ETFs, the top 10 ETFs with the highest 5-day average "Net Active Buy Amount Proportion" include Southern SSE STAR Chip ETF (27.5%, 100.0% percentile) and E Fund Hang Seng Dividend Low Volatility ETF (23.6%, 99.6% percentile) [16]

产业观察 [table_Header]2025.12.02 【具身智能产业动态】松延动力完成近 2 亿元 Pre- 产业研究中心 B 轮融资，灵心巧手完成数亿元 A 加轮融资 摘要：产业最新趋势跟踪，点评产业最新风向 [Table_Summary] 具身机器人产业动态 国内规模最大的全尺寸人形机器人数据集 LET 数据集开源发布 工信部人形机器人标准化技术委员会名单公布，王兴兴、稚晖君入选 极佳视界发布轮式人形机器人 Maker H01 智元机器人上线灵心平台，用户可通过简易操作实现个性化机器人配置 星动纪元发布物流仓储解决方案，全球首次端到端 VLA 在物流仓储真实应用 智能汽车产业动态 汽车销量榜：10月新势力交付排行榜前三名为零跑（70289辆）、鸿蒙智行（68216 辆）、小鹏（42013 辆）。 小马智行发布 Q3 财报，城市级 UE 转正，Robotaxi 千台目标将提前完成，2026 年规模扩至 3000 辆 小马智行与阳光出行建立战略合作，全面推进自动驾驶出行业务发展 九识智能与广州高新兴机器人战略合作，共推无人巡检创新应用 西井科技与和记阿曼苏哈尔港口签约，15 台 E-Truck 即将进驻 ...

Investment Rating - The industry investment rating is "Overweight" [6] Core Insights - On December 1, ByteDance's Doubao team released a technical preview of the Doubao mobile assistant, integrating the Doubao large model deeply into the operating system, advancing edge AI from application-level to system-level [3][7] - The Doubao mobile assistant allows users to interact through voice, side buttons, or the Doubao Ola Friend headset, enhancing user experience and interaction [7] - The Doubao large model demonstrates international first-class performance in reasoning, visual understanding, image creation, video generation, and voice capabilities, achieving the best results in multiple authoritative evaluations [7] Summary by Sections Industry Overview - The report focuses on the integration of AI technology into mobile systems, highlighting the collaboration between Doubao and mobile manufacturers to create a system-level service [3][7] Investment Highlights - The Doubao mobile assistant is positioned as a core product for mobile manufacturers, facilitating the broadest application scenarios and commercial monetization paths for large models [7] - The assistant is currently available for developers and tech enthusiasts on the engineering sample nubia M153, priced at 3499 yuan, with plans for ecological cooperation with various mobile brands [7]