🎉Community Raffle - Win $25

An exclusive raffle opportunity for active members like you! Complete your profile, answer questions and get your first accepted badge to enter the raffle.
Join and Win

Output of Neural Net operator is more than Linear Regression Output

User: "nschauh"
New Altair Community Member
Updated by Jocelyn

Hello,

 

I have a question here, I have used winequality-red training and scoring datasets in RapidMiner (with quality as the target attribute) and applied cross validation (Apply model and performance operator) with first Linear regression operator and replaced it with Neural Net Operator with number of folds =2

 

The output for Linear Regression is : root_mean_squared_error: 0.652 +/- 0.003 (mikro: 0.652 +/- 0.000)

and 

The output for Neural Net Operator is : root_mean_squared_error: 0.741 +/- 0.031 (mikro: 0.741 +/- 0.000)

 

Is this a right prediction that I have performed and what can be inferred from both the root mean squared errors?

From my understanding, for this dataset the performance of the Linear Regression is better as a predictive model than the Neural Net predictive model. Can that be the case?

 

I look forward for your response please.

 

Kind Regards,

Namrata

 

Sort by:
1 - 1 of 11
    User: "Thomas_Ott"
    New Altair Community Member

    Yes it could very much be the case that the Linear Regression is slightly better than the Neural Net.