Advances, Systems and Applications
From: Cost-effective land cover classification for remote sensing images
Notation | Defination |
---|---|
m | The mth iteration during the clustering process |
x | A d-dimensional data point |
\(x_i\) | The ith d-dimensional data point in cluster x |
\(c_j\) | The d-dimensional center of the cluster j |
\(u_{ij}\) | The degree of membership of \(x_j\) in the cluster j |
\(\mathcal {J}_m\) | The objective function at the mth iteration |
\(P_i\) | The ith partition during the clustering process |
\(Rand(P_1,P_2)\) | The Rand Index of two partitions \(P_1\) and \(P_2\) |
\(\mathcal {L}_m\) | The predicted labels at the mth iteration |
\(r_m\) | The accuracy at the m iteration |
\(\Delta {\mathcal {J}_m}\) | The change rate of \(\mathcal {J}_m\) |
\(\bar{r}\) | The desired accuracy (e.g., 85%, 90%, 95%, etc.) |
\(\Delta \bar{\mathcal {J}}\) | The predicted objective function given the \(\bar{r}\) |
\(\text {Cost}_{\text {comp}}\) | The computation cost in the cloud |
\(\text {Price}_{\text {unit}}\) | The unit price |
\({T}_{\text {comp}}\) | Total computation time |
\({T}_{\text {train}}\) | The time taken in the training process |
\(T_{\text {actual}}\) | The early-stop computation time in the clustering |
\({T}_\textit{total}\) | The computation time for achieving 100% accuracy |
\(\text {Cost}_{\text {effective}}\) | The cost-effectiveness percentage |