\begin{figure}[h] \centering \includegraphics[width=0.8\textwidth]{your-image-file.png} \caption{The distribution of average temporal absolute errors.} \label{fig:errors} \end{figure}
As shown in Fig. 9, each method's performance in error variability is significantly different. Specifically, the PCA and SRSVD methods have the most significant error fluctuation amplitude and degree, showing high instability. In contrast, the FWA method's error fluctuation range is relatively small, which indicates that it has higher stability and consistency. This stability and consistency are essential for netTM estimates, as they directly affect the accuracy and reliability of network management and optimization.