行业轮动周报：预先调整下大盘很难再现四月波动，融资资金净流出通信-20251013

• The report introduces the Diffusion Index Model for industry rotation, which has been tracked for four years. The model is based on momentum strategies to capture industry trends. It showed strong performance in 2021 with excess returns exceeding 25% before experiencing a significant drawdown due to cyclical stock adjustments. In 2022, the strategy delivered stable returns with an annual excess return of 6.12%. However, in 2023 and 2024, the model faced challenges, with annual excess returns of -4.58% and -5.82%, respectively. For October 2025, the model suggests allocating to industries such as non-ferrous metals, banking, communication, steel, electronics, and automobiles[26][30] - The Diffusion Index Model is constructed by ranking industries based on their diffusion index values, which reflect upward trends. The top six industries as of October 10, 2025, are non-ferrous metals (0.98), banking (0.951), communication (0.909), steel (0.877), electronics (0.823), and automobiles (0.813). The bottom six industries are food and beverage (0.137), consumer services (0.297), real estate (0.407), coal (0.445), transportation (0.457), and construction (0.489)[27][28][29] - The Diffusion Index Model achieved an average weekly return of 2.59%, exceeding the equal-weighted return of the CSI First-Level Industry Index by 0.70%. For October, the model's excess return is -0.37%, while the year-to-date excess return is 4.60%[30] - The report also discusses the GRU Factor Model for industry rotation, which utilizes minute-level price and volume data processed through a GRU deep learning network. The model has shown strong adaptability in short cycles but struggles in long cycles and extreme market conditions. Since February 2025, the model has focused on growth industries but has faced difficulties in capturing excess returns due to concentrated market themes[32][38] - The GRU Factor Model ranks industries based on GRU factor values. As of October 10, 2025, the top six industries are comprehensive (6.64), building materials (5.21), construction (3.55), textile and apparel (3.31), transportation (2.99), and steel (2.88). The bottom six industries are computing (-41.87), food and beverage (-35.34), electronics (-34.87), non-ferrous metals (-28.25), power equipment and new energy (-26.61), and communication (-22.71)[33][36] - The GRU Factor Model achieved an average weekly return of 2.88%, exceeding the equal-weighted return of the CSI First-Level Industry Index by 1.01%. For October, the model's excess return is 1.67%, while the year-to-date excess return is -6.55%[36]