ASOS GraphReturns
收藏资源简介:
ASOS GraphReturns数据集由大型在线时尚零售商ASOS创建,旨在解决预测客户退货的挑战。该数据集包含约290万条记录,涵盖2021年9月至11月的购买和退货数据,涉及约77万至82万独特客户和约41万产品变体。数据集自然形成图结构,适用于图表示学习,特别是用于开发图神经网络模型以提高退货预测的准确性。该数据集的应用领域主要集中在通过预测退货来优化客户体验和减少零售成本,同时支持可持续发展目标。
The ASOS GraphReturns dataset was developed by ASOS, a large global online fashion retailer, to tackle the challenge of customer return prediction. It contains approximately 2.9 million records covering purchase and return data from September to November 2021, involving roughly 770,000 to 820,000 unique customers and about 410,000 product variants. The dataset naturally exhibits a graph structure, making it well-suited for graph representation learning, especially for building graph neural network (GNN) models to enhance the accuracy of return prediction. Its main application scenarios focus on optimizing customer experience and reducing retail operational costs through return prediction, while also supporting sustainable development goals.

- 1A Dataset for Learning Graph Representations to Predict Customer Returns in Fashion Retail伦敦大学学院 · 2023年



