100% Accuracy?

Hello everyone, I am new to Knime.
I was doing a simple exercise: prediction of the number of bikes used during hours, and I used the Linear Regression method to do this. I obtained an accuracy of 100%, every prediction is correct, also if I decrease the number of training samples to only 1%. Can you say me what I am doing wrong? Thanks.
exercise.knwf (19.1 KB)

I tried to edit the post, but I didn’t find any “Edit” button, sorry.

I wanted to add to the previous post that I tried also using a decision tree learner and I obtained “normal results” not 100% accuracy. How is it possible that linear regression classified all the rows in the correct way?
exercise.knwf (28.3 KB)

Please post the data set as well.

1 Like

The datasets are here:

A linear regression is for regression problems. Accuracy is used to measure the performance of classification tasks. You could rather focus on MSE or RMSE

1 Like

Hi @CasellaJr

It is because you are feeding the answer to the machine. You set your target column to “cnt_y” but if you check the data you’ll see "cnt_y column is simply sum of “causual_y” and “registered_y” columns!
So you should use column filter node and filter these two columns.



Yeah, thank you @mehrdad_bgh I found this solution, I removed that variables from the algorithm! Thank you


This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.