Loading...

Performance Evaluation of Tag Recommendation in Online Social Networking Q & A

Khezrian, Navid | 2021

360 Viewed
  1. Type of Document: M.Sc. Thesis
  2. Language: Farsi
  3. Document No: 54071 (19)
  4. University: Sharif University of Technology
  5. Department: Computer Engineering
  6. Advisor(s): Habibi, Jafar
  7. Abstract:
  8. Open-source online social networks for project sharing and online social networks Q&A make use of tags and keywords for indexing, classification, and thematic search. users are responsible for selecting tags to identify their content, which can lead to human errors, or malicious users with improper tagging can lead to information corruption. Uncontrolled use of words also leads to the production of different types of tags, leading to redundancy or ambiguity. The most obvious advantage of tagging is the correct classification of information, which provides better services for relevant searches and requests. In this study, we propose a new model called TagBERT, which for the first time uses deep learning and BERT pre-training to recommend tags in open-source online social networks for project sharing and online social networks Q&A. In this model, the processed sentences are first converted into numerical vectors by means of the tokenizer of BERT. Next, the attributes are extracted using the CNN network and, afterwards, the DNN network is trained on the extracted attributes in order to recommend tags. To evaluate our model, we used four datasets, i.e. Free-code, UNIX, Wordpress, and Software Engineering. Our proposed model obtained the highest precision score over baseline deep-learning and conventional methods. In contrast to previous studies in which precision was significantly reduced as a result of increased recommended tags, the precision of our model did not remarkably vary with increase in the number of tags
  9. Keywords:
  10. Classification ; Bidirectional Encoder Representations from Transformers (BERT)Model ; Question-Answer Sites ; Online Question-Answer Communities ; Open-Source Communities ; Tag Recommendation

 Digital Object List

 Bookmark

No TOC