经过归一化的多次实验的 TikZ 直方图

Question

我想绘制多个实验的直方图，其中可能事件是 1 到 10 之间的数字。由于实验的抽取次数不同，因此直方图应显示标准化为 100 次抽取的每个实验。

为了便于输入，此规范化应由宏执行\expandrow，该宏将总绘制次数和一个数字列表作为参数，该列表指定观察到特定事件的频率。例如，\expandrow{31}{5,4,1,9,2,8,0,0,2}= (1, 16.12854) (2, 12.90283) (3, 3.22571) (4, 29.03137) (5, 6.45142) (6, 25.80566) (7, 0.0) (8, 0.0) (9, 6.45142)。宏可以独立运行，但如果在\addplot命令中应用则不起作用。我该如何解决这个问题？

  \documentclass{standalone}
    \usepackage{tikz}
    \usepackage{pgfplots}
    \pgfplotsset{compat=1.12}

    \newcommand{\expandrow}[2]{
          \foreach \v [count = \i] in {#2} {
              (\i, \pgfmathparse{\v/#1 * 100}\pgfmathresult)
        }
    }

  \begin{document}

      \begin{tikzpicture}
      \begin{axis}[
      ybar,
      enlarge y limits={abs=2em},
      bar width = {.4em},
      enlargelimits=0.15,
      ylabel={percentage},
      symbolic x coords={1, 2, 3, 4, 5, 6, 7, 8, 9, 10},
      legend style={at={(0.5,-0.15)},
          anchor=north,legend columns=-1},
      nodes near coords = {\tiny \pgfmathprintnumber\pgfplotspointmeta},
      nodes near coords align={vertical},
      ]
      % A
      % \addplot coordinates {\expandrow{31}{5,4,1,9,2,8,0,0,2}};
      \addplot coordinates {
          (1, 16.12854) (2, 12.90283) (3, 3.22571) (4, 29.03137) (5, 6.45142) (6, 25.80566) (7, 0.0) (8, 0.0) (9, 6.45142)
      };
      % B
      % \addplot coordinates {\expandrow{1739}{157,1285,246,47,4}};
      \addplot coordinates {
          (1, 9.0271) (2, 73.89221) (3, 14.1449) (4, 2.70233) (5, 0.22888)
      };        

      \legend{A, B}

      \end{axis}
      \end{tikzpicture}

  \end{document}

Answer 1

如果执行额外的步骤，并首先将规范化的坐标列表存储在宏中，那么您可以执行类似以下代码的操作，我借用了 egreg 对 是否可以存储由 foreach 生成的数组？

\documentclass{standalone}
\usepackage{pgfplots}
\pgfplotsset{compat=1.12}
\makeatletter
\newcommand\normalize[3]{%
  \gdef\store@temp{}%
  \foreach [count=\i] \v in {#2}{\pgfmathparse{\v/#1 * 100}\xdef\store@temp{\store@temp (\i,\pgfmathresult)}}%
  \let#3=\store@temp
}
\makeatother

\begin{document}

  \begin{tikzpicture}
  \begin{axis}[
  ybar,
  enlarge y limits={abs=2em},
  bar width = {.4em},
  enlargelimits=0.15,
  ylabel={percentage},
  symbolic x coords={1, 2, 3, 4, 5, 6, 7, 8, 9, 10},
  legend style={at={(0.5,-0.15)},
      anchor=north,legend columns=-1},
  nodes near coords = {\tiny \pgfmathprintnumber\pgfplotspointmeta},
  nodes near coords align={vertical},
  ]
  % do the normalization
  \normalize{31}{5,4,1,9,2,8,0,0,2}{\rowA}
  \normalize{1739}{157,1285,246,47,4}{\rowB}

  % plot the data
  % A
  \addplot coordinates {\rowA};

  % B
  \addplot coordinates {\rowB};

  \legend{A, B}

  \end{axis}
  \end{tikzpicture}
\end{document}