arXiv:2602.07653v2 Announce Type: replace Abstract: The tensor-train (TT) format is a data-sparse tensor representation commonly used in high dimensional data approximations. In order to represent data with interpretability in data science, researchers develop data-centric skeletonized low rank approximations. However, these methods might still suffer from accuracy degradation, nonrobustness, and high computation costs. In this paper, given existing skeletonized TT approximations, we propose a family of projection enhanced interpolation based algorithms to further improve approximation accuracy while keeping low computational complexity. We do this as a postprocessing step to existing interpolative decompositions, via oversampling data not in skeletons to include more information and selecting subsets of pivots for faster projections. We provide a priori error bounds for the projection enhanced TT format and for the proposed dimension parallel algorithm, in which the contributions of the skeleton, the oversampling, and the oblique projections appear as separate terms. We illustrate the performances of our proposed methods with extensive numerical experiments. These include up to 10D synthetic datasets such as tensors generated from kernel functions, and 4D and 6D tensors constructed from Maxwellian distribution functions that arise in kinetic theory. Our results demonstrate accuracy improvements between 2 and 100 times on tested tensors over original skeletonized TT approximations, while using a limited amount of computational resources.