[Objective] This paper proposes a supplementary question and answer (Q&A) recommendation method based on the transfer learning enhanced multi-label multi-document Q&A classification model, aiming to identify and recommend supplementary Q&A in community question answering.
[Methods] We proposed new features and classified Q&A supplementary relationships combining existing features. Then, we built a transfer learning enhanced multi-label multi-document Q&A classification model to identify and recommend supplementary Q&As.
[Results] Experiments are conducted on three meta-tasks using real datasets. Results demonstrate that the proposed method improves 48.3%, 15.8%, and 32.5% in the precision, recall, and f-measure separately.
[Limitations] The method has only been applied to health Q&A topics on the Zhihu platform. The performance on different platforms with different topics of Q&A remains to be verified.
[Conclusions] The proposed recommendation method can effectively recommend supplementary Q&As. It helps users in community question answering acquire knowledge more comprehensively and promote knowledge utilization in the community.
李莹, 李明.
基于迁移学习增强的多标签多文档分类模型的补充性问答推荐研究
[J]. 数据分析与知识发现, 10.11925/infotech.2096-3467.2023.0683.
Li Ying, Li Ming.
Supplementary Q&A recommendation based on transfer learning enhanced multi-label multi-document classifier
. Data Analysis and Knowledge Discovery, 0, (): 1-.