Step $1$: Read and explore the data, split the data into training and test

We use the same dataset as we used for Week $2$ Exercise. You will have the same steps before we create a logistic regresion model.

$[]$

Start coding or generate with AI$.$

Assign variables to X and y$,$ perform feature scaling, and split the data into training and test set, using $25\%$ of the data as test set. Make sure set random$\_$state $=0$ as we did in Week $2$ Decision Tree Exercise. In this way we created the same training set and test set, so we can compare the model performance of decision tree and logistic regression.

$[]$

Start coding or generate with AI$.$

Step $2$: Create a Logistic Regression Model

Create a logistic regression model named logreg using LogisticRegression$().$

What is the model accuracy on the training set? Hint: you may use $.$score$()$ to get the model accuracy, covered in Week $2$ Lab.

$[]$

Then we evalute the model performance on the test set and see how it generalizes to a dataset that is never used to create the model.

First get the predictions for test set using $.$predict$(),$ save them in y$\_$pred, then use accuracy$\_$score$()$ to get the model accuracy on test set.

What is model performance on the test set?

$[]$

Start coding or generate with AI$.$

Now let's create a dataframe to show the coefficient of each variable in this logistic regression model. We first save the names of all variables in to an object called heart$\_$feature$\_$names, then use it as the index in $.$DataFrame$().$

The coefficients are saved in the first element in $.$coef$\_$ attribute.

Which variable has the highest coefficient?

$[]$

Start coding or generate with AI$.$

Then we estimate the probability of having heart disease for each data point on the test set $.$ We use $.$predict$\_$proba$().$

What is the probability of second row in test set to have heart disease $($look at the second row and second column in the output$)?$

$[]$