扩散指数择时模型
Search documents
金工专题:基于相对强弱视角下的扩散指数择时模型
China Post Securities· 2025-06-25 10:46
- The report introduces two models: the first model is based on the diffusion index within a single index to judge the style rotation, and the second model is based on the relative strength between different indices for rotation[2][4][5] - The first model involves dividing the constituent stocks of an index into two groups based on a specific factor, calculating the proportion of stocks in each group that outperform the average return over a past period, and then computing the relative strength indicator[16][18] - The second model extends the first model by comparing two different indices, calculating the weighted average return of the two indices, and then computing the relative strength indicator between the two indices[26][30] - The backtesting of the first model using the double moving average method shows that the optimal look-back period is around 20-25 days, and the model performs well in terms of stability[6][31] - The backtesting of the second model shows that the model can achieve relatively stable results with a look-back period of around 20 days[6][42] - The parameter sensitivity analysis indicates that the optimal parameters for the first model are a look-back period of 20-25 days, and for the second model, the optimal parameters are a look-back period of 20-30 days and around 180 days[49][50][54][57]