如何在TraMineR熵图中添加回归线？

Question

我想知道是否以某种方式将线性回归线添加到TraMineR熵图中。

示例：

Private Function CreatePath(ByVal area As Rectangle) As GraphicsPath
    '* Set fill mode to winding (see FillMode Help)
    Dim gp As New GraphicsPath(FillMode.Winding)

    Dim rect As Rectangle = New Rectangle(area.Location, New Size(area.Width, area.Height \ 3))
    gp.AddEllipse(rect)

    '* Reorder points so that path goes clockwise (see FillMode.Winding Help)

    Dim hadj As Integer = area.Height \ 4
    'Left side bottom -> top
    'gp.AddLine(New Point(area.X, area.Y + area.Height - hadj), New Point(area.X, area.Y + (rect.Height \ 3) + 8))

    '* Right side top -> bottom (clockwise)
    '* NOTE: Top Y value was a bit too high, adjusted by -3.  Adjust calculation or maybe use floating point math and used rounding, ceil, or floor
    gp.AddLine(New PointF(area.X + area.Width, area.Y + (rect.Height \ 3) + 9 - 3), New PointF(area.X + area.Width, area.Y + area.Height - hadj))

    Dim gh As Integer = area.Width \ 4
    Dim pts(4) As PointF
    'pts(0) = New PointF(area.X, area.Y + area.Height - hadj)
    'pts(1) = New Point(area.X + gh, area.Y + area.Height - (hadj \ 2))
    'pts(2) = New Point(area.X + gh * 2, area.Y + area.Height - (hadj \ 3))
    'pts(3) = New Point(area.X + gh * 3, area.Y + area.Height - (hadj \ 2))
    'pts(4) = New Point(area.X + area.Width, area.Y + area.Height - hadj)
    '* Reordered points (clockwise)
    pts(0) = New Point(area.X + area.Width, area.Y + area.Height - hadj)
    pts(1) = New Point(area.X + gh * 3, area.Y + area.Height - (hadj \ 2))
    pts(2) = New Point(area.X + gh * 2, area.Y + area.Height - (hadj \ 3))
    pts(3) = New Point(area.X + gh, area.Y + area.Height - (hadj \ 2))
    pts(4) = New PointF(area.X, area.Y + area.Height - hadj)
    gp.AddCurve(pts)

    'Right side top -> bottom
    'gp.AddLine(New PointF(area.X + area.Width, area.Y + (rect.Height \ 3) + 9), New PointF(area.X + area.Width, area.Y + area.Height - hadj))

    '* Left side bottom -> top (clockwise)
    '* NOTE: Top Y value was a bit too high, adjusted by -2.  Adjust calculation or maybe use floating point math and used rounding, ceil, or floor
    gp.AddLine(New Point(area.X, area.Y + area.Height - hadj), New Point(area.X, area.Y + (rect.Height \ 3) + 8 - 2))

    Return gp
End Function

我想要这样的东西：

我知道我可以用library(TraMineR) data(biofam) set.seed(10) biofam <- biofam[sample(nrow(biofam),300),] biofam.lab <- c("Parent", "Left", "Married", "Left+Marr", "Child", "Left+Child", "Left+Marr+Child", "Divorced") biofam.seq <- seqdef(biofam, 10:25, labels=biofam.lab) seqplot(biofam.seq, type="Ht")

计算熵

seqient()

但是由于数据是宽格式的，所以我不确定如何获取数据，以便执行biofam$ient <- seqient(biofam.seq, with.missing=TRUE) 方法。

abline(lm(y ~ x))

> head(biofam[c(1, 10:12, 28)]) idhous ... a15 a16 a17 ... Entropy 1234 NA ... 0 0 0 ... 0.3010904 1515 86261 ... 0 0 0 ... 0.3154649 276 17561 ... 0 0 0 ... 0.4012026 1212 69021 ... 0 0 0 ... 0.5517478 153 11391 ... 0 0 0 ... 0.2559298 1164 66641 ... 0 0 0 ... 0.4178207 似乎也没有提供像seqplot()这样的数值矢量/矩阵：

plot()

Answer 1

由seqHtplot或seqplot(..., type="Ht")渲染的截面熵存储在$Entropy返回的对象的seqstatd元素中。每个时间点的值是该时间点的截面状态分布的熵。因此，每个时间点只有一个值。

这与seqient返回的纵向熵不同。每个序列有一个值。该值是序列内状态分布的熵。

因此，要绘制回归线，您只需要使用seqstatd提取横截面熵，定义时间变量，然后在abline(lm())中使用这些变量：

stat.bf <- seqstatd(biofam.seq)
ent <- stat.bf$Entropy
time <- 1:length(ent)

seqplot(biofam.seq, type = "Ht")
abline(lm(ent ~ time), col = "red", lty = 2)