*5.4. Comparison of Multi-Step Prediction Results*

Moreover, in order to evaluate the influence of proposed PGPM with the input time sequences of various step lengths, an experiment was also implemented based on different time steps, and the experimental results are shown in Table 7.

From Table 8, it can be found that there was a positive correlation between the prediction error and step size; in other words, the prediction error increased with respect to step length increases. Synchronously, the fitting accuracy had a negative correlation with step length, that is, the fitting accuracy decreased as the step length increased. The reason for the above phenomenon is that the dependence between the power generation and time sequences is weakened with the increase of step length. In summary, when the time step of input time sequences is four, the PGPM proposed in this paper can meet the demand for power generation forecasting.


**Table 8.** Comparison of multi-step prediction results.
