Reproduce “CoGSeq-LTV: Scalable Missing-Aware Graph-Sequence Modeling for Multi-Business Transaction Analytics” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ziyang Su, Yao Zhou, Kelin Liu, Guanjie Zheng, Linghe Kong
CoGSeq-LTV predicts users’ cumulative 7-, 14-, 30-, and 45-day spending in e-commerce and food delivery. It builds user–item graphs within mini-batches, encodes each business’s transaction history separately, and aligns same-user representations across businesses. It achieves the best reported results on all eight targets; e-commerce 7-day MAE is 261.55 ± 39.43. Predictions are not constrained to increase across horizons.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.