classification learner app accuracy changed with every run command why?

3 次查看（过去 30 天）

显示更早的评论

geetu yadav 2016-5-24

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/285710-classification-learner-app-accuracy-changed-with-every-run-command-why

评论： Walter Roberson 2016-5-25

same feature dataset its accuracy changed for every classifier evrytime why? how can i get correct result?

回答（1 个）

Walter Roberson 2016-5-24

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/285710-classification-learner-app-accuracy-changed-with-every-run-command-why#answer_223330

I would suspect that some of the work involves random numbers. If you use rng() to set the same seed each time before classification then you would probably get the same result each time.

2 个评论
显示无隐藏无

geetu yadav 2016-5-25

how to do that pls explain while using learner app?

Walter Roberson 2016-5-25

在 MATLAB Online 中打开

Call

rng(12345)

then invoke the learner app and train once. Then exit the learner app and repeat the sequence, rng(12345) and go back into the learner app and train once again. I suspect you will get the same results.

请先登录，再进行评论。

请先登录，再回答此问题。