How to create a binary svm classifier?

2017 7 18
2 个回答
更新时间：2017 7 18
1 次查看（30 天）

0 个投票

I have 5 sets of data each with 3 runs of eeg signal acquisitions, and I'm to create a classifier that will distinguish between the rest and mental activity in each run. So first ten seconds are mental activity and then rest for 20s. The class labels are given but its just [1 2 1 2 1 2 1 2 1 2 1 2] for class 1 being mental activity and 2 being rest, for all runs and subjects? To use fitcsvm I need the Y labels array but am not sure how to get that

请先登录，再回答此问题。

回答（2 个）

Walter Roberson
Walter Roberson 2017-7-18

0 个投票

The Y labels vector is the same as the class labels vector.

4 个评论
显示 2更早的评论 隐藏 2更早的评论

Somaia Ahmadi
Somaia Ahmadi 2017-7-18
This was given along with the data; does it change after filtering or putting all the training signals into a single matrix?
Walter Roberson
Walter Roberson 2017-7-18
Do any of those operations change the column's identification as belonging to mental activity or rest?
Somaia Ahmadi
Somaia Ahmadi 2017-7-18
编辑：Somaia Ahmadi 2017-7-18
Nope it should still be mental activity or rest, but I'm worried that if I put all the runs from each subjects together in one matrix the class labels won't apply..
Walter Roberson
Walter Roberson 2017-7-18
As long as you keep columns separate you will be fine. (Each column is an individual sample.)

请先登录，再进行评论。

Image Analyst
Image Analyst 2017-7-18

0 个投票

Did you try the Classification Learner app, on the Apps tab, and have the wizard export the code for you?

1 个评论
显示 -1更早的评论 隐藏 -1更早的评论

Somaia Ahmadi
Somaia Ahmadi 2017-7-18
Yeah I did, but it wasn't able to find the response variable (which I assume is the class label vectors in the cell array containing the signals)

请先登录，再进行评论。

类别

在 帮助中心 和 File Exchange 中查找有关 Classification Learner App 的更多信息

产品

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by