TeX 62.3%
Python 37.6%
1\begin{table}[htbp]2\centering\scriptsize3\caption{Computational cost (wall-clock on a 32-core Mac Studio M3 Ultra; tree models 10 threads, neural models 8 threads)}\label{tab:cost}4\begin{threeparttable}5\begin{tabular}{lrrrrr}6\toprule7Task & Model & Seconds per fit & Total seconds & Fits & Mean training rows \\8\midrule9A & Ridge & 0.1 & 1 & 12 & 314{,}363 \\10A & Hedonic OLS & 0.3 & 4 & 12 & 314{,}363 \\11A & Elastic net & 3.4 & 40 & 12 & 314{,}363 \\12A & Quantile MLP & 7.7 & 93 & 12 & 314{,}363 \\13A & HSHN (full pooling) & 9.0 & 108 & 12 & 314{,}363 \\14A & HSHN (no pooling) & 9.3 & 111 & 12 & 314{,}363 \\15A & HSHN (partial pooling) & 9.5 & 114 & 12 & 314{,}363 \\16A & XGBoost & 9.9 & 119 & 12 & 314{,}363 \\17A & Random forest & 29.9 & 359 & 12 & 314{,}363 \\18A & SVR (subsample) & 63.6 & 764 & 12 & 314{,}363 \\19A & LightGBM & 155.0 & 1860 & 12 & 314{,}363 \\20B & No change & 0.0 & 0 & 36 & 5{,}785 \\21B & AR(2), cell & 0.0 & 0 & 36 & 5{,}785 \\22B & Seasonal naive & 0.0 & 0 & 36 & 5{,}785 \\23B & Mean growth & 0.0 & 0 & 36 & 5{,}785 \\24B & Ridge (panel) & 0.0 & 0 & 36 & 5{,}785 \\25B & Random walk with drift & 0.0 & 0 & 36 & 5{,}785 \\26B & Momentum & 0.0 & 0 & 36 & 5{,}785 \\27B & BHPF & 0.2 & 7 & 36 & 5{,}785 \\28B & QANSS & 0.2 & 8 & 36 & 5{,}785 \\29B & MLP (panel) & 1.0 & 36 & 36 & 5{,}785 \\30B & Local linear trend & 1.6 & 58 & 36 & 5{,}785 \\31B & RAREF & 3.9 & 140 & 36 & 5{,}785 \\32B & DSGFN & 4.9 & 176 & 36 & 5{,}785 \\33B & TFT-lite & 5.0 & 181 & 36 & 5{,}785 \\34B & LightGBM (panel) & 5.8 & 210 & 36 & 5{,}785 \\35B & GRU & 7.1 & 254 & 36 & 5{,}785 \\36B & LSTM & 8.9 & 322 & 36 & 5{,}785 \\37B & QREFA & 10.4 & 373 & 36 & 5{,}785 \\38B & TCN & 186.0 & 6696 & 36 & 5{,}785 \\39\bottomrule40\end{tabular}41\end{threeparttable}42\end{table}43