Mismatch in the dimensions

#1

I am not able to understand why I am getting this mismatch in the dimensions. Here is the drive file and the screenshot of the error

drive.google.com/file/d/1Z554WjUl1AgpO6w5OCJwhGB9ZdUk36nJ/view?usp=sharing

1347×584 21.4 KB

#2

hey @A18ML0031 ,
problem in your code is that .
You adding an extra 1 in X_train, but not doing so for X_test.
So due to this , your X_train is of shape (,4) where as X_test is of shape (,3).
Just do that for X_test also.
Everything will work absolutely fine.

I hope this helped you.
Thank You and Happy Learning :slightly_smiling_face:.

#3

Ohhh. I didn’t catch that. Thanks for helping out.

#4

My accuracy is coming out to be 29% now. Should i share the notebook file?

#5

Yeah please it would easy for me to help you out then.

#6

Here is the drive link to the file: drive.google. com/file/d/1Z554WjUl1AgpO6w5OCJwhGB9ZdUk36nJ/view?usp=sharing

#7

Remove the space after ‘google’ and before ‘.com’

#8

hey @A18ML0031 ,
if you check your code , then you have mistakenly taken different variable to store weights while training your model. ( small w and capital W )

So just change them to be same then you will get very good results.

#9

The ‘W’ I have used is to indicate that I am initialising the random weights which will give the loss, weights and bias from the gradient_descent method used to train the model. However, further when I am predicting the results for X_test, I am using the weights that have been learned by the model on training on the X_train based on the random weights ‘W’. The final weights are stored in the variable ‘w’ which I have used when making predictions.

#10

Are you saying that I am supposed to keep the variables same to be either “W” or “w” throughout the code?

#11

Yes.
Because we update our weights and for that we need to use the same variable of weights , all over same.
As with that only we will be able to update our weights and train properly.

#12

Hi, I have made the required changes. I am still getting very low accuracy after all that

#13

do one thing , kindly update your code and upload it on google colab and share its link here.
Let me have a look at it once.