An improved TDA-based algorithm for the isomorphism problem of orthogonal arrays

In previous work, we introduced a topology-based classification algorithm for the isomorphism check of orthogonal arrays. Here, we discuss some theoretical properties of the isomorphisms, and how they impact our classifier. We then leverage these properties to give an improved version of the algorithm, and discuss its performance under various parameter choices. In so doing, we address an open conjecture of ours. We conclude by comparing its complexity with respect to an established exact approach, and discussing its range of applications.