Using Machine Learning to Detect Exoplanets in Outer Space

In this chapter, we shall learn how to detect exoplanets in outer space using ensemble methods that are based on decision trees.

Decision trees are a family of non-parametric supervised learning methods. In a decision tree algorithm, the data is divided into two partitions by using a simple rule. The rule is applied again and again to further partition the data, thus forming a tree of decisions.

Ensemble methods combine the learning from multiple learning algorithms to improve predictions and reduce errors. These ensembles are differentiated on the basis of what kind of learners they use and how they structure those learns in the ensemble.

The two most popular ensemble methods based on decision trees are known as gradient boosted trees and random forests. 

The following topics will be covered in this chapter:

  • What is a decision tree?
  • Why we need ensembles?
  • Decision tree-based ensemble methods
    • Random forests
    • Gradient boosting
  • Decision tree-based ensembles in TensorFlow
  • Building a TensorFlow boosted tree model for exoplanet detection
The code from this chapter is available in Jupyter Notebook as  ch-02_Detecting_Explonaets_in_Outer_Space.ipynb in the code bundle.