In a world full of Data Science booming, there are so many different possibilities with Machine Learning. Choosing which one to prefer for "your" strategic project is difficult. So many possible flavors and combinations, it can get confusing . Does not help that the outcome of our model will be based on the propitious of the methodology choosen.
No Pressure...
Here is my go to reference guide to what packages/methodologies I should look at. This is great because one project you might be looking at building a recommendation engine, while the other is classification model of behavioral insights due to a promotion. Picking the right methodology is a must, spend time on this moment of the build
More details can be found here- Great interactive guide http://scikit-learn.org/stable/tutorial/machine_learning_map/
An alternative without broken down to classification, clustering, dimensionality reduction, and regression is this table. I have found it very useful especially with population sizes.
There you have it, later I will post some examples of my work using these various methods. Specifically for my background in behavioral sciences I have evolved from originally doing K-means clustering to provide offers, or k-means to predict fraudulent behavioral for the government now I use agglomeration Bayesian clustering (ABC).
If your interested in reading more about agglomeration Bayesian clustering method this is a great study. https://www.siam.org/meetings/sdm06/proceedings/044wellingm.pdf
Thanks for Reading! Feel free to comment