given the Classified Data
with entirely unknown features, we face a challenge of not having any background/expertise knowledge on the given target class. via elementary exploratory data analysis on these unknown variables, we realised the target variable is a binary outcome while all of its other predictors are continuous variables.
employing k Nearest Neighbours (kNN) classification after scaling the data will the primary method to execute this prediction.
to find out the optimal k-value which most accurately predict the target classes (0 or 1) via kNN classification
- initial exploratory data analysis
- kNN classification with k=1
- choose optimal k-value from for-loop
- kNN classification with optimal k-value
- comparative analysis between initial & optimal k-value via confusion matrix and classification report