This
Data Challenge
is to put everything you learned as part of Supervised Machine Learning Module into practice. The dataset we will be using contains two CSV files 'features' and ‘target’.
Each row in the
‘features’
belongs to a
‘measurement’
and each column represents a
‘feature’
. For each row in the
‘features’
you have a corresponding class label in
‘target’
. You can consider the row-numbers as keys.
You will have to use data labelled as
train
for implementing the models. You have to submit the results that you obtain from the predictions of your Machine Learning models from the data labelled as
test
.
Complete the following:
1. Explore your data. Can you identify anything interesting that is worth noting from the data?
2. Define a set of possible classifiers and show which one performs best. Keep in mind the problem of overfitting.
3. Using feature selection try to reduce the number of features. In the dataset, you have about 120 features. Find the good ones for your classifier.

Keep in mind that there is
no best solution
to the challenge. Show how you approach a problem, and the skills/methods you use in that approach.