Title | ||
---|---|---|
A Tag-Based Transformer Community Question Answering Learning-to-Rank Model in the Home Improvement Domain |
Abstract | ||
---|---|---|
Community Question Answering (CQA) is an Information Retrieval (IR) task that allows matching complex subjective questions and candidate answers based on user posts in community web forums. User questions and comment-based answers deal with many problems, such as redundancy or ambiguity of linguistic information. In this paper, we propose a pairwise learning-to-rank model community QA model in the home improvement domain. For a user question, this model must rank candidate answers in order of relevance. Our main contribution consists of transformer-based language models using user tags to accurate the model generalisation. To train our model, we also propose a proper CQA dataset in home improvement domain that consists of information extracted from community forums. We evaluate our approach by comparing the performance based on analysis with the state-of-the-art method on text or document similarity. |
Year | DOI | Venue |
---|---|---|
2021 | 10.1007/978-3-030-86475-0_13 | DATABASE AND EXPERT SYSTEMS APPLICATIONS, DEXA 2021, PT II |
Keywords | DocType | Volume |
Information retrieval, Community question answering, Neural networks | Conference | 12924 |
ISSN | Citations | PageRank |
0302-9743 | 0 | 0.34 |
References | Authors | |
0 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Macedo Maia | 1 | 6 | 3.15 |
Siegfried Handschuh | 2 | 1988 | 181.71 |
Markus Endres | 3 | 1 | 2.38 |