Instructions

Read through this blog and the Logistic$\_$Regression.ipynb file that comes

with this task as guides to applying Logistic Regression.

Read this article from Datacamp to learn about preprocessing categorical

variables.

Compulsory Task $1$

Follow these steps:

Begin by reading the Iris data set $($Iris$.$csv$)$ into a Jupyter notebook and

name it iris$\_$logistic$\_$regression.ipynb.

The dataset consists of three classes of irises. The objective is to create a

classifier that will predict whether an iris belongs to the 'Iris$-$setosa' class

or not.

This means that we have two classes: 'Iris$-$setosa' and not$-$'Iris$-$setosa'

$($which includes 'Iris$-$versicolor' and 'Iris$-$virginica'$).$

Identify your independent variables $x.$

Encode your dependent variable $y$ such that 'Iris$-$setosa' is encoded

as $0,$ and 'Iris$-$versicolor' and 'Iris$-$virginica' are both encoded as

$(0$ corresponds to the 'Iris$-$setosa' class, and $1$ corresponds to the

not$-$'Iris$-$setosa' class.$)$

Split the data into a training and test set.

Use sklearn's logistic regression function to fit a model and make

predictions on the test set.

Use sklearn to generate a confusion matrix, which compares the

predicted labels to the actual labels $($gold labels$).$

Analyse the confusion matrix and provide a prediction, in a comment, whether the model is likely to have hgher precison, higher recall or similar precison and recall. Write your own code to calulate the accuracy, precision and recall, and check whether your prediction was right