Check out the first Dataiku 8 Deep Dive focusing on Productivity on October 29th Read More

train test errors & Hyperparameters tuning

Level 1
train test errors & Hyperparameters tuning

Hi,

I've made a model (say Linear Regression) using a visual recipe and randomly split my data into 0.8 and 0.2 ratio for train and validation/test data. Now, after the model has been trained, I get different error metrics, R2 Score, scatter plot, and distribution of errors. I have the following questions-

1)  These analyses are on the train data, not on validation/test data? 

2) How can I check the error metrics on test data if the above-shown metrics are on the train data?

3) If I want to use a few more training parameters for modeling than it is available for any model by default, is there any plugin to do that visually?

Note: I could write a pyspark query and get all things done but the aim here is to deliver the solution with visually available recipes, which is much more friendly to use.

 

Thanks,

Nitish

0 Kudos
1 Reply
Dataiker
Dataiker

Hi @nitish 

As answered on your subsequent post, all of these questions are clarified in our docs and training materials. Since your question covers several stages of the ML process, I recommend that you have a look at the materials in question here

Good luck!

0 Kudos