Description
I have an uploaded document that has all the details.
Unformatted Attachment Preview
Note: put your code in functions and use default arguments as makes sense.
You are now expected to use at least some functions in your code to make it
reusable. We will check that there is at least one function in your code when grading.
● the credit approval: dataset where you are determining if someone will be
approved for a credit card.
● Q1- Clean your dataset to turn categorical values into numerical ones. Onehot encoding is likely the answer, but it depends on the dataset. Your data
may have ordinal columns, for example where one-hot encoding is not as
appropriate.
● Q2- Perform univariate linear regression on the dataset. Select your variable
to predict. How well did this model perform? Is this a good approach for this
dataset? Why or why not?
● Q3- Perform KNN on this dataset. As part of this, write a function that selects
the optimal value of k. How well did this model perform?
● Q4- Work with your dataset to perform logistic regression. How well did this
perform?
● Q5- Perform normalization on your dataset. Does it change the performance
for 2-4? What is the best measure of performance for your dataset (accuracy
or something else) and why?
● Q6- Use type hints in your functions to make sure you are using the right
types when you call the functions. https://realpython.com/lessons/type-hinting/
Purchase answer to see full
attachment
