Integer optimization practices, with ~60 constrains. Real fun with actual implementations. With that, I declare the wrap up of this AMAZING course :)

Playing with kmeans clustering and trying to group customers based on their behaviours (unsupervised learning)

Another encore --> Done with the 2nd part of the exam Having fun and a lot of trials with CART and Cross-validation to calculate model accuracy.

Finished writing codes + answering question of the 1st exam in this amazing course!

A bit similar to the previous week but using constrained (integers) as variables. Fun, and am very satisfied for the past 9 weeks. =))

Practiced and learned a lot to use linear optimization in spreadsheet software. Most important are: - defining problems and variables - setting up correct constraints and objectives

Gentle intros and exercises to: - maps, merge - ggplot - wordcloud - data analysis and plots modification and arts :))

Learned 2 most clustering techniques: - Hierarchical Clustering with dendograms plot + scree plot - kmeans Clustering, by defining centroids

Possibly one of the most computationally demanding weeks at MIT analytics edge so far. Lessons learned: - Using 3 machine learning models (CART, logistic regression and random Forest) to predict outcomes - Played a bit with text analytics (natural language comprehension) - bear witness the powerful IBM Watson

This week is a blast! I learned how to use: - logistic regression - regression trees - random forest and tuning parameters to improve models' accuracy! =))

Three separate homework for week 3: all experiments and some codes to finish the job!

I finished 3 main tasks in this assignment. The core part is using linear regression model to predict dependent variables and to test how independent variables affect the strength of our model. Learned a lot!

Some reruns of the simulations, after and while watching the lectures!

This is the solution for Analytics EDGE MITx Course Feel free to study but not distribute it elsewhere

First week lessons recap

My first trial with interactive table that can be downloaded as csv or excel file