Book Image

Hands-On Data Science and Python Machine Learning

By : Frank Kane
Book Image

Hands-On Data Science and Python Machine Learning

By: Frank Kane

Overview of this book

Join Frank Kane, who worked on Amazon and IMDb’s machine learning algorithms, as he guides you on your first steps into the world of data science. Hands-On Data Science and Python Machine Learning gives you the tools that you need to understand and explore the core topics in the field, and the confidence and practice to build and analyze your own machine learning models. With the help of interesting and easy-to-follow practical examples, Frank Kane explains potentially complex topics such as Bayesian methods and K-means clustering in a way that anybody can understand them. Based on Frank’s successful data science course, Hands-On Data Science and Python Machine Learning empowers you to conduct data analysis and perform efficient machine learning using Python. Let Frank help you unearth the value in your data using the various data mining and data analysis techniques available in Python, and to develop efficient predictive models to predict future results. You will also learn how to perform large-scale machine learning on Big Data using Apache Spark. The book covers preparing your data for analysis, training machine learning models, and visualizing the final data analysis.
Table of Contents (11 chapters)

Decision Trees in Spark with MLlib

Alright, let's actually build some decision trees using Spark and the MLlib library, this is very cool stuff. Wherever you put the course materials for this book, I want you to go to that folder now. Make sure you're completely closed out of Canopy, or whatever environment you're using for Python development, because I want to make sure you're starting it from this directory, OK? And find the SparkDecisionTree script, and double-click that to open up Canopy:

Now, up until this point we've been using IPython notebooks for our code, but you can't really use those very well with Spark. With Spark scripts, you need to actually submit them to the Spark infrastructure and run them in a very special way, and we'll see how that works shortly.

...