Implement Naive Bayes Classifier from scratch to predict if a person makes over 50K a year. Again, you cant use an API that implements Naive Bayes, but implementing intermediate functions as API is fine.
dataset: https://archive.ics.uci.edu/ml/datasets/adult
Compute: Precision, recall, accuracy (you can use APIs for these functions)
Process: do a hold out method of your dataset with the following split
60-40, 70-30, 80-20 (training-testing)
What you need to submit:
a. Code
b. a document describing your approach as well as three graphs,
-
Y- axis: precision, x axis: hold out split
-
Y- axis: recall, x axis: hold out split
-
Y- axis: accuracy, x axis: hold out split