Diabetes Prediction using KNN

[vanshika230]

Workflow

1. Data Analysis
2. Applying KNN
3. Saving Predictions Please assign me this issue under Hacktoberfest.

[Anubhavdevv]

Yup sure go ahead @vanshika230

[void-02]

Hello, I am willing to contribute to this issue.

[Anubhavdevv]

Sure, @void-02 Go Ahead

[void-02]

Hello @anubhav201241, please let me know the dataset which is to be used.

[Anubhavdevv]

Hey, @void-02 You can use any dataset, you can go with knn after that you can shift to ANN for more accuracy !!

[void-02]

Hello @anubhav201241 , I have tried both KNN and MLP on a diabetes dataset, but I found the accuracy to be decreasing in case of MLP. Here are the details- KNN:- accuracy when n-neighbors were 3 = 68.83 accuracy when n-neighbors were 5 = 71.42 accuracy when n-neighbors were 7 = 72.72 accuracy when n-neighbors were 10 = 72.73 accuracy on best_parameter = 74.02

MLP:- accuracy when hidden layer size was 50 = 67.53 accuracy when hidden layer size was 100 = 63.63 accuracy when hidden layer size was 150 = 64.93 accuracy when hidden layer size was 200 = 67.53

Should I create a new pull request ?

[Anubhavdevv]

Hey, @void-02 Thanks for trying out yup accuracy is a bit low so you can try a different algorithm and there is no need to create another pull req just commit in your file and push that code0 to your repo then it will also show in your PR