市场情绪监控周报（20250519-20250523）：本周热度变化最大行业为综合、交通运输-20250526

• Model name: Broad-based index rotation strategy; Model construction idea: Using the weekly change rate of the total heat of different broad-based indices to construct a rotation strategy; Model construction process: Calculate the weekly heat change rate of different groups of stocks and smooth it using MA2. On the last trading day of each week, buy the broad-based index with the highest total heat change rate MA2. If the highest change rate is in the "other" group, stay in cash. The strategy's net value is shown in the report. The formula is: $ \text{Heat Change Rate MA2} = \frac{\sum_{i=1}^{n} \text{Heat Change Rate}i}{n} $ where $ n $ is the number of weeks. The strategy's annualized return since 2017 is 8.74%, with a maximum drawdown of 23.5%. The return since 2025 is 8.13%[14][16][17] - Model name: Broad-based index heat; Model construction idea: Aggregate the total heat of component stocks in broad-based indices; Model construction process: Group the entire A-share sample according to the component stocks of the CSI 300, CSI 500, CSI 1000, CSI 2000, and "other" groups. Sum the total heat indicators of the component stocks in each group to obtain the heat of the four mainstream broad-based indices and the heat of stocks not included in these indices ("other" group). The formula is: $ \text{Total Heat} = \sum{i=1}^{n} (\text{Browse Count}_i + \text{Self-selection Count}_i + \text{Click Count}i) $ where $ n $ is the number of stocks in the group. The heat is normalized by dividing by the total market heat and multiplying by 10000, with a range of [0,10000][9][12] - Factor name: Total heat indicator; Factor construction idea: Aggregate the browsing, self-selection, and click counts of stocks; Factor construction process: Define the total heat indicator as the sum of the browsing, self-selection, and click counts of a stock, normalized by its proportion in the entire market on the same day, and multiplied by 10000. The formula is: $ \text{Total Heat} = \frac{(\text{Browse Count} + \text{Self-selection Count} + \text{Click Count})}{\text{Total Market Heat}} \times 10000 $ with a range of [0,10000][8] - Factor name: Concept heat; Factor construction idea: Aggregate the total heat of component stocks in different concepts; Factor construction process: Calculate the weekly heat change rate of each concept and smooth it using MA2. Construct two simple portfolios: select the top 5 concepts with the highest heat change rate each week, exclude the smallest 20% of stocks by market cap, and equally hold the top 10 stocks with the highest total heat in each concept to form the heat TOP portfolio. Similarly, select the bottom 10 stocks with the lowest total heat in each concept to form the BOTTOM portfolio. The formula is: $ \text{Heat Change Rate MA2} = \frac{\sum{i=1}^{n} \text{Heat Change Rate}_i}{n} $ where $ n $ is the number of weeks. The BOTTOM portfolio's annualized return is 15.71%, with a maximum drawdown of 28.89%. The return since 2025 is 18.35%[28][30][32][33] Model backtest results - Broad-based index rotation strategy, annualized return 8.74%, maximum drawdown 23.5%, return since 2025 8.13%[17] - Broad-based index heat, CSI 300 heat change rate increased by 5.25%, CSI 500 heat change rate decreased by 13.72%[17] Factor backtest results - Total heat indicator, normalized range [0,10000][8] - Concept heat, BOTTOM portfolio annualized return 15.71%, maximum drawdown 28.89%, return since 2025 18.35%[33]