Abstract
Random Forests are one of the most popular classifiers in machine learning. The larger they are, the more precise the outcome of their predictions. However, this comes at a cost: it is increasingly difficult to understand why a Random Forest made a specific choice, and its running time for classification grows linearly with the size (number of trees). In this paper, we propose a method to aggregate large Random Forests into a single, semantically equivalent decision diagram which has the following two effects: (1) minimal, sufficient explanations for Random Forest-based classifications can be obtained by means of a simple three step reduction, and (2) the running time is radically improved. In fact, our experiments on various popular datasets show speed-ups of several orders of magnitude, while, at the same time, also significantly reducing the size of the required data structure.
Highlights
Random1 Forests are one of the most widely known classifiers in machine learning [2,19]
We present an optimisation method that is based on algebraic aggregation: Random Forests are transformed into a single decision diagram in a semanticspreserving fashion, which, in particular, preserves the learner’s variance and accuracy
We have presented an approach to aggregate large Random Forests into single and compact decision diagrams that faithfully reflects the semantics of the original Random Forest for a considered purpose
Summary
Random Forests are one of the most widely known classifiers in machine learning [2,19]. We present an optimisation method that is based on algebraic aggregation: Random Forests are transformed into a single decision diagram in a semanticspreserving fashion, which, in particular, preserves the learner’s variance and accuracy. The great advantage of the resulting decision diagrams is their absence of redundancy: during classification every predicate is considered at most once, and only if its evaluation is required This allows one to obtain concise explanations and evaluation times that are optimal.. Key to our approach are Algebraic Decision Diagrams (ADDs) [28] Their algebraic structure supports compositional aggregation, abstraction, and reduction operations that lead to minimal normal forms. Using basic algebraic operations, such as concatenation and addition, allows us to aggregate a Random Forest into a single ADD that faithfully maintains the individual results of each tree in the forest. Abstracting results (i.e. the leaf structure of the decision diagrams) to the essence, in this case the
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callMore From: International Journal on Software Tools for Technology Transfer
Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
