Welcome to Monika's Machine Learning repository. The intend is to showcase some of my favorite projects:
-
Classification problem:
-
I have shown two examples under this category.
-- Example I: This is a very simple case of recommending best airlines based on the number of cancellations it has had in the past.
-- Example II: This is a more complex example of classification type Machine Learning problem simply because there is a severe class imbalance in the data set. I have demonstrated that in the first iteration I am not able to achieve good metrics for any of the models used. So I completely change the methodology and start from scratch. Second time around, I train my models based on equal number of failure and non-failure cases which renders much improved confusion matrix scores.
-
-
Natural Language Processing:
-- Shows Machine Learning models used for Natural Language Processing (NLP). Here numerous text messages are identified as ham or spam.
-- Analyzed Urdu data set from UC Berkeley repository
-
A/B testing:
-- Used Kaggle dataset to show A/B testing using the frequentist and Bayesian method.