Back to Search Start Over

Automatic Literature Mapping Selection: Classification of Papers on Industry Productivity

Authors :
Guilherme Dantas Bispo
Guilherme Fay Vergara
Gabriela Mayumi Saiki
Patrícia Helena dos Santos Martins
Jaqueline Gutierri Coelho
Gabriel Arquelau Pimenta Rodrigues
Matheus Noschang de Oliveira
Letícia Rezende Mosquéra
Vinícius Pereira Gonçalves
Clovis Neumann
André Luiz Marques Serrano
Source :
Applied Sciences, Vol 14, Iss 9, p 3679 (2024)
Publication Year :
2024
Publisher :
MDPI AG, 2024.

Abstract

The academic community has witnessed a notable increase in paper publications, whereby the rapid pace at which modern society seeks information underscores the critical need for literature mapping. This study introduces an innovative automatic model for categorizing articles by subject matter using Machine Learning (ML) algorithms for classification and category labeling, alongside a proposed ranking method called SSS (Scientific Significance Score) and using Z-score to select the finest papers. This paper’s use case concerns industry productivity. The key findings include the following: (1) The Decision Tree model demonstrated superior performance with an accuracy rate of 75% in classifying articles within the productivity and industry theme. (2) Through a ranking methodology based on citation count and publication date, it identified the finest papers. (3) Recent publications with higher citation counts achieved better scores. (4) The model’s sensitivity to outliers underscores the importance of addressing database imbalances, necessitating caution during training by excluding biased categories. These findings not only advance the utilization of ML models for paper classification but also lay a foundation for further research into productivity within the industry, exploring themes such as artificial intelligence, efficiency, industry 4.0, innovation, and sustainability.

Details

Language :
English
ISSN :
20763417
Volume :
14
Issue :
9
Database :
Directory of Open Access Journals
Journal :
Applied Sciences
Publication Type :
Academic Journal
Accession number :
edsdoj.1796551f28b4e688650d5fcb5a49c30
Document Type :
article
Full Text :
https://doi.org/10.3390/app14093679