Prediction based on best fit linear regression model

Question

Mekala balaji 2017-8-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/352319-prediction-based-on-best-fit-linear-regression-model

评论： the cyclist 2017-8-11

在 MATLAB Online 中打开

Hi,

I have the below data(train_data, test_data), and I want to do following:

In train set some of point or outliears, and I want evaluate best fit model based on mean squared error
Prediction based on best fit model.

Train_data:
x   y
1   1
2   3
6   6
8   2
11  15
15  9
21  11
25  14
Test set:
x 
5
12
21
23
my desired output(i.e, y:)
4
7
12
13

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

the cyclist 2017-8-11

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/352319-prediction-based-on-best-fit-linear-regression-model#answer_277712

It looks like you have the Statistics and Machine Learning Toolbox. I would use fitlm to fit the model on the training data.

Then you can use the predict method to make a prediction on the test set.

2 个评论
显示无隐藏无

Mekala balaji 2017-8-11

Sir, I can use fitlm. But sir, there are some outliers due which the rsquare is very poor, I want to evaluate by taking how many input data (drop others) such that we get higher rsquare.

the cyclist 2017-8-11

You can do "robust" fitting (with fitlm, or other MATLAB functions). This is a common way of handling outliers. The documentation page I linked has the details on how to do this.

This Wikipedia page discusses robust methods.

请先登录，再进行评论。