Relation-aware Meta-learning for Market Segment Demand Prediction with Limited Records

Recently, E-commerce platforms have extensive impacts on our human life. To provide an efficient platform, one of the most fundamental problem is how to balance the demand and supply in market segments. While conventional machine learning models have achieved a great success on data-sufficient segments, it may fail in a large-portion of segments in E-commerce platforms, where there are not sufficient records to learn well-trained models. In this paper, we tackle this problem in the context of market segment demand prediction. The goal is to facilitate the learning process in the target segments even facing a shortage of related training data by leveraging the learned knowledge from data-sufficient source segments. Specifically, we propose a novel algorithm, RMLDP, to incorporate a multi-pattern fusion network (MPFN) with a meta-learning paradigm. The multi-pattern fusion network considers both local and global temporal patterns for segment demand prediction. In the meta-learning paradigm, the transferable knowledge is regarded as the model parameter initializations of MPFN, which are learned from diverse source segments. Furthermore, we capture the segment relations by combining data-driven segment representation and segment knowledge graph representation and tailor the segment-specific relations to customize transferable model parameter initializations. Thus, even with limited data, the target segment can quickly find the most relevant transferred knowledge and adapt to the optimal parameters. Extensive experiments are conducted on two large-scale industrial datasets. The results show that our RMLDP outperforms a set of state-of-the-art baselines. In addition, RMLDP has also been deployed in Taobao, a real-world E-commerce platform. The online A/B testing results further demonstrate the practicality of RMLDP.