Doubt in diabetes classification challenge

#1

I am getting 73% accuracy in the diabetes classification challenge. The link to my code is - https://drive.google.com/drive/folders/1EVNFeA5WrdG7OV5_Qa0Bh1T9NDNsc6xN?usp=sharing . Please tell what to do

#2

hey @abhaygarg2001 ,
Your code is really good.
Some things you can perform to increase your accuracy are:

  1. Trying different values of K , as its value depend upon the dataset.
  2. Trying scaling the data , by using different scaling methods available in sklearn module.
  3. Averaging predictions of 2 or more KNN models having different values of K.

With KNN , you can at max achieve a score of 77-79% ,not more than that and if you did then it would really be appreciable.

If you want to achieve accuracy around 85-95% then you should different models and techniques , for example bagging ,boosting , voting ,etc. With proper hyper parameter tuning you can achieve great results.

I hope this would have resolved your doubt.
Thank You.
Happy Coding. :slightly_smiling_face: