TeX 62.3%
Python 37.6%
1\begin{table}[htbp]2\centering\footnotesize3\caption{Leakage audit}\label{tab:leakage}4\begin{threeparttable}5\begin{tabular}{p{4cm}p{11cm}}6\toprule7Item & Treatment \\8\midrule9Hedonic coefficients & WP22 stage 1 re-estimated at every vintage; Task A features never use coefficients \\10Geographic clustering / location cells & WP22 nested grid rebuilt from data $\le$ origin at every vintage; Task A uses raw coordinates \\11Normalisation & standardisation constants from training rows only; index levels enter as changes \\12Feature selection & fixed feature lists, no selection on test data \\13Imputation & WP22 conditional medians (no price information); indicators time-stable \\14Target encoding & none; one-hot with training categories \\15Spatial graph & DSGFN adjacency recomputed at each origin from centroids and the real-time history $\le$ origin \\16Hyper-parameter tuning & fixed a priori (Table~\ref{tab:hyper}) \\17Index revisions & predictors = vintage published for the origin; target = final series (revision component reported) \\18Partial last month & October 2026 excluded; last target September 2026 \\19\bottomrule20\end{tabular}21\end{threeparttable}22\end{table}23