Logistic regression fit line plot

4 次查看（过去 30 天）

Asma 2016-2-16

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/268463-logistic-regression-fit-line-plot

评论： Kevin Claytor 2016-2-16

testData.csv

I am trying to find the logistic regression between one independent variable and one dependent variable. And then to plot a regression line (on scatter plot). My dependent variable is continuous (i.e. no categories). My question is can I plot the line for such type of data. For reference my data is also attached where 1st column is independent variable. And the code I am using is as below: (But it gives me a flat line). Any ideas or example links?

[b,dev,stats] = glmfit(x,[y1 repmat(100,size(y1))], 'binomial', 'link', 'logit')
xx = linspace(min(x), max(x), 50);
yfit = glmval(b,xx,'logit');
figure(2),plot(x,y1,'o',xx,yfit,'-')

1 个评论
显示 -1更早的评论隐藏 -1更早的评论

Kevin Claytor 2016-2-16

Why do you pad your data in glmfit - what does 'repmat(100,size(y1))' do? To me, this generates a flat line at a level of 100. If I were to fit this, the logistic curve should be flat.

请先登录，再进行评论。

请先登录，再回答此问题。