High-performance FAQ retrieval using an automatic clustering method of query logs

Automated FAQ answering with question-specific knowledge representation for web self-service

“…In this sense our technique correlates with [12] (see Section IV.A). Nonetheless, the data in a template is not statistical.…”

“…[12] categorizes previously logged queries using original FAQs as the seed data. The system matches a new query not only to the original FAQs but also to the attached clusters of previous queries, which ensures larger lexical diversity.…”

“…In a nutshell, MRR value is 1 if all queries get the right answer as the first item in the list of retrieved texts, and 0.5 if the right answer is always the second retrieved text. The best MRR for the answer retrieval component of [12] was 0.7, given that all other components of the system made no mistakes.…”

A high-performance FAQ retrieval method using minimal differentiator expressions

“…Auto-FAQ [36] works in a similar way, a keyword comparison criterion is applied to implement the question matching. Others methods like [37] use automatic clustering of the registered questions in order to increase the number of question to have more variability in the examples. International conferences like Text REtrieval Conference (TREC) [38] have promoted the development of new techniques that mix statistical methods and NLP.…”

Question Similarity Calculation for FAQ Answering

“…Berger et al [5] proposed a statistical lexicon correlation method. User click log has also been used to find similar queries in [6].…”