Accuracy only 41% Improvement required

what should be done to improve accuracy,
how can i improve the accuracy using sklearn.

I know that we should not use sklearn for assignments but i first tried with custom svm implementation and got only 39 % accuracy i tried fixing learning rate and maxItr in custom and C and max_iter in sklearn yet i coudnt get it better.

i request you to please guide me how to get maximum accuracy possible custom implementation code https://ide.codingblocks.com/s/330025

hey @S19LPN0163 ,
i checked both of your code links , they are really good and well implemented.
Now coming to the part of accuracy , i would suggest you to try with different models.
Its not that you can only use SVM, you can can any other model too.
Just try that once , and i hope it will work better.

Thank You and Happy Learning :slightly_smiling_face:.

but in one dicussions i saw vasu bhaiya told it is possible to achieve accuracy of more than 90% ans moreover if svm itself is giving this much accuracy then wont knn be better than svm for classification of more than 2 classes

while working with images , conventional machine learning algorithms doesn’t so good.
Try with neural networks , you will get much better accuracy.

ya i understand that but the thing is I want to know that is it only me who is getting this much accuracy or its the same with everybody. If possible could you tell me your accuracy using svm in this challenge. I am worried about accuracy as here vasu bhaiya told it is possible to achieve > 90% accuracy for this particular challenge Image classification using a linear classifier (SVM Pokemon Project)

heres the link of the discussion if required Image classification using a linear classifier (SVM Pokemon Project)

yeah i saw the discussion.
I too tried once with SVM but could achieve a max acc. of 40% , i don’t know how vasu sir is saying about it as such. But i think , with SVM this is the max you can achieve on this dataset.

Hope this helps you.

ok thank you very much

I guess your doubt is now resolved.
So i would request you to kindly mark it as resolved in your course’s doubt section and do provide your valuable feedback as it would help us to improve this platform and provide better learning experience.

Thank You and Happy Learning :slightly_smiling_face:.