论文写作 17: Latex 表格的控制

1. 表格一般应该与文字边距相同. 如果表格内容不多, 可以用如下语句控制宽度

\begin{tabular*}{12cm}{@{\extraclosep{\fill}}lll}

2. 如果数据太多装不下, 也可以用如下语句控制宽度, 但表格中字体会自动缩小.

\resizebox{12cm}{!}{

它应将整个的 tabular 括起来.

3. 对齐方面, l 表示 left, c 表示 center, r 表示 right.
   应该根据所投期刊发表论文的风格来调整对齐方式. 在多数情况下, 小数点位置应该对齐. 例如, 某一列里面有整数 30 与 35,420 之类, 就应该使用 r; 而 0.32 与 0.453 就应该使用 l.

范例如下.

\begin{table}[h]\caption{Dataset information}\label{table: datasets}
    \centering
    \setlength{\tabcolsep}{12pt}
    \begin{tabular*}{12cm}{@{\extracolsep{\fill}}lrrlll}
    \hline
       Dataset & $|\mathbf{U}|$  & $|\mathbf{C}|$ &      Area & $rt$ & $qi$\\
    \hline
         Seeds &             210 &              8 &      Life & 0.19 & 0.1\\
       Thyroid &             215 &              6 &      Life & 0.25 & 0.1\\
         Heart &             270 &             14 &      Life & 0.05 & 0.1\\
        Spiral &             312 &              3 & Synthetic & 0.3  & 0.1\\
    Ionosphere &             350 &             35 &  Physical & 0.04 & 0.1\\
           R15 &             600 &              3 & Synthetic & 0.2  & 0.1\\
    USps\_2\_6 &           2,200 &            257 &     Image & 0.2  & 0.1\\
      Waveform &           5,000 &             22 &  Physical & 0.1  & 0.1\\
        Credit &           5,987 &             66 & Financial & 0.35 & 0.1\\
       Twonorm &           7,400 &             21 & Synthetic & 0.3  & 0.1 \\
    \hline
    \end{tabular*}
\end{table}

效果如图 1 所示.

图 1. 表格例 1

\begin{table*}[!htb]
	\caption{Accuracy comparison with MAL algorithms. $-$ indicates unable to obtain due to high time complexity.}
	\label{table: comparison-multi-label-active}
	\centering
    \resizebox{\textwidth}{!}{
    \begin{tabular}{lcccccc}
        \hline
        Dataset       & AUDI'13 & AURO'15 & CS-GP'19 & GP-B$^2$M'21 & CMAL'21 & MASP               \\
        \hline
   		Art           & $.9275_{\pm .0039}$ & $.9392_{\pm .0012}$ & $.9371_{\pm .0006}$ & $.9367_{\pm .0012}$ & $.9372_{\pm .0015}$   & $.9374_{\pm .0013}$ \\
    	Birds         & $.9412_{\pm .0070}$ & $.9457_{\pm .0063}$ & $.9028_{\pm .0170}$ & $.9357_{\pm .0151}$ & $.9468_{\pm .0058}$   & $.9469_{\pm .0053}$ \\
    	Business      & $.9680_{\pm .0029}$ & $.9715_{\pm .0012}$ & $.9471_{\pm .0009}$ & $.9673_{\pm .0066}$ & $.9710_{\pm .0019}$   & $.9718_{\pm .0016 }$ \\
    	Cal500        & $.7826_{\pm .0179}$ & $.8199_{\pm .0131}$ & $.8503_{\pm .0019}$ & $.8503_{\pm .0039}$ & $.8516_{\pm .0068}$   & $.8367_{\pm .0070}$ \\
    	Computer      & $.9545_{\pm .0023}$ & $.9610_{\pm .0008}$ & $.9543_{\pm .0017}$ & $.9544_{\pm .0011}$ & $.9615_{\pm .0013}$   & $.9619_{\pm .0015}$ \\
    	Education     & $.9491_{\pm .0018}$ & $.9573_{\pm .0016}$ & $.9556_{\pm .0008}$ & $.9544_{\pm .0016}$ & $.9559_{\pm .0007}$   & $.9557_{\pm .0004 }$ \\
    	Emotion       & $.6875_{\pm .0209}$ & $.6808_{\pm .0213}$ & $.6886_{\pm .0091}$ & $.6886_{\pm .0185}$ & $.7039_{\pm .0179}$   & $.7425_{\pm .0157 }$ \\
    	Enron         & $.8792_{\pm .0391}$ & $.9331_{\pm .0083}$ & $.9363_{\pm .0012}$ & $.9235_{\pm .0064}$ & $.9408_{\pm .0016}$   & $.9443_{\pm .0036}$ \\
    	Entertain     & $.9238_{\pm .0037}$ & $.9382_{\pm .0033}$ & $.9324_{\pm .0011}$ & $.9322_{\pm .0020}$ & $.9362_{\pm .0026}$   & $.9378_{\pm .0031}$ \\
    	Flags         & $.5444_{\pm .1016}$ & $.6086_{\pm .0556}$ & $.5153_{\pm .0136}$ & $.4942_{\pm .0172}$ & $.6603_{\pm .0368}$   & $.6327_{\pm .0280}$ \\
    	Genbase       & $.9182_{\pm .0259}$ & $.9487_{\pm .0150}$ & $.9536_{\pm .0009}$ & $.9536_{\pm .0015}$ & $.9574_{\pm .0050}$   & $.9536_{\pm .0018}$ \\
    	Health        & $.9502_{\pm .0042}$ & $.9591_{\pm .0016}$ & $.9481_{\pm .0006}$ & $.9469_{\pm .0031}$ & $.9540_{\pm .0016}$   & $.9562_{\pm .0008}$ \\
    	Image         & $.6263_{\pm .0736}$ & $.6609_{\pm .0518}$ & $.7528_{\pm .0032}$ & $.7529_{\pm .0032}$ & $.7668_{\pm .0093}$   & $.7612_{\pm .0079}$ \\
    	Recreation    & $.9291_{\pm .0015}$ & $.9362_{\pm .0016}$ & $.9352_{\pm .0026}$ & $.9349_{\pm .0014}$ & $.9365_{\pm .0012}$   & $.9334_{\pm .0020}$ \\
    	Reference     & $.9641_{\pm .0012}$ & $.9711_{\pm .0005}$ & $.9648_{\pm .0008}$ & $.9610_{\pm .0014}$ & $.9688_{\pm .0023}$   & $.9701_{\pm .0009}$ \\
    	Scene         & $.7543_{\pm .0294}$ & $.7155_{\pm .0124}$ & $.8210_{\pm .0033}$ & $.8210_{\pm .0009}$ & $.8429_{\pm .0063}$   & $.8601_{\pm .0066}$ \\
    	Science       & $.9552_{\pm .0051}$ & $.9645_{\pm .0006}$ & $.9637_{\pm .0010}$ & $.9634_{\pm .0008}$ & $.9640_{\pm .0008}$   & $.9647_{\pm .0005}$ \\
    	Social        & $.9716_{\pm .0014}$ & $.9780_{\pm .0009}$ & $.9673_{\pm .0006}$ & $.9582_{\pm .0061}$ & $.9755_{\pm .0008}$   & $.9759_{\pm .0006}$ \\
    	Society       & $.9310_{\pm .0035}$ & $.9428_{\pm .0015}$ & $.9373_{\pm .0013}$ & $.9367_{\pm .0022}$ & $.9400_{\pm .0026}$   & $.9420_{\pm .0020}$ \\
    	Yeast         & $.6588_{\pm .0317}$ & $.6757_{\pm .0558}$ & $.6973_{\pm .0008}$ & $.7569_{\pm .0086}$ & $.7749_{\pm .0088}$   & $.7698_{\pm .0075}$ \\
        {\color{yellow}Bibtex }       & $-$ & $.9813_{\pm .0035}$ & $.9847_{\pm .0004}$ & $.9846_{\pm .0006}$ & $.9860_{\pm .0002}$   & $.9849_{\pm .0006}$ \\
    	{\color{yellow}Corel5k }      & $-$ & $.9896_{\pm .0001}$ & $.9905_{\pm .0001}$ & $.9899_{\pm .0005}$ & $.9903_{\pm .0001}$   & $.9906_{\pm .0001}$ \\
        \hline
        Mean rank  & 5.25 & 2.9 & 4.175 & 4.525 & 2.1 & 2.05\\
        \hline
    \end{tabular}}
\end{table*}

效果如图 2 所示.

图 2. 表格例 2

说明如下:

• 所有列都为左对齐, 使用 0.0000 而不是 0 来保证对齐.
• $\pm$ 用于显示标准差. 使用下标以节约版面.
• 粗体指示最好的效果, 斜体指示次好的效果.
• Meanrank 指示不同数据集的总体效果.
• 数据集应按照某种方式排序, 如对象数、属性数、来源领域等, 并保持所有表格中的顺序一致.
• 大于 1000 的数据应该用逗号隔开, 如 1,024.
• 作为良好的习惯, 源码里面也应该对齐, 而不要因为不影响最终 pdf 效果而偷懒.