Tree-based models have proven to be remarkably useful for prediction tasks in machine learning and have recently been applied to and modified for problems in economics and finance. The fundamental unit of any tree-based model is a decision tree, which explains an outcome using a sequence of data partitions. Such a model can be naturally visualized as a flowchart.
While TensorFlow was developed for the purpose of solving deep learning problems, it has recently added libraries for tree-based models in its high-level Estimators API. In this chapter, we’ll examine those libraries and will ...
