AN AGENT-BASED DOCUMENT CLASSIFICATION MODEL TO IMPROVE THE EFFICIENCY OF THE AUTOMATED SYSTEMATIC REVIEW PROCESS
This paper proposes an Agent-based Document Classification (AbDC) model that computerizes the systematic literature review (SLR) process by imitating what a researcher is supposed to perform during the literature review process manually. The AbDC model comprises three main components that perform th...
Saved in:
| Main Authors: | , , |
|---|---|
| Other Authors: | |
| Format: | Review |
| Published: |
Little Lion Scientific
2023
|
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This paper proposes an Agent-based Document Classification (AbDC) model that computerizes the systematic literature review (SLR) process by imitating what a researcher is supposed to perform during the literature review process manually. The AbDC model comprises three main components that perform the SLR. Firstly, the document classification algorithm analyses a full text of research articles and evaluates relevancy. Secondly, the multi-agent architecture accelerates the mining process and handles the performance issues. Finally, the web-based systematic review tool tests and validates the functionality of the proposed AbDC model. The first testing was conducted to assess the performance of the proposed AbDC. Result shows that the required processing time was reduced by 33.5% using four agents to achieve the mining process. Meanwhile, the second testing was performed to validate the mining process results. The text extraction method was run on 200 documents from various studies to conduct the review process. The parsing process yielded valid results with 98.5% accuracy. The testing results showed that the proposed AbDC model is significant in providing researchers and postgraduate students with new means to perform SLR. � 2022 Little Lion Scientific |
|---|