Vouw: Geometric Pattern Mining using the MDL Principle
This work addresses a novel pattern mining problem for researchers in data mining and machine learning, focusing on geometric data analysis.
The paper tackles the problem of discovering recurring local structures in discrete geometric matrices by introducing geometric pattern mining, which identifies complex spatial relations and arbitrarily shaped patterns, and demonstrates that their Vouw algorithm achieves high-quality results on a synthetic benchmark.
We introduce geometric pattern mining, the problem of finding recurring local structure in discrete, geometric matrices. It differs from existing pattern mining problems by identifying complex spatial relations between elements, resulting in arbitrarily shaped patterns. After we formalise this new type of pattern mining, we propose an approach to selecting a set of patterns using the Minimum Description Length principle. We demonstrate the potential of our approach by introducing Vouw, a heuristic algorithm for mining exact geometric patterns. We show that Vouw delivers high-quality results with a synthetic benchmark.