Scholarlink

Writing a thesis can be challenging for university students, mainly due to the need for specific language tools in Spanish. This study presents an online platform that assists the thesis writing process by analyzing lexical richness, using Variety, Density, and Sophistication. Additionally, the platform includes a section to identify the structure of objectives and answer critical questions about what is going to be done? (Q1), What will it be done for? (Q2) And how will this action be carried out? (Q3). This platform aims to improve the quality of university theses. During the development, a usability survey was conducted with students, achieving an average rating of 8.12 out of 10, which allowed for identifying and correcting details for improvement. One of our work’s contributions is using current web technologies combined with pre-trained BERT models, which will enable quick and easy user analysis.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Enhancing Student Theses with Advanced Text Analysis Using NLP and Pre-trained Models

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Maximiliano Ponce Marquez,
  • Samuel González-López,
  • Jesús Raúl Cruz Rentería,
  • Gilberto Borrego Soto,
  • Manuel Omar Meranza Castillon,
  • Guillermina Muñoz Zamora

摘要

Writing a thesis can be challenging for university students, mainly due to the need for specific language tools in Spanish. This study presents an online platform that assists the thesis writing process by analyzing lexical richness, using Variety, Density, and Sophistication. Additionally, the platform includes a section to identify the structure of objectives and answer critical questions about what is going to be done? (Q1), What will it be done for? (Q2) And how will this action be carried out? (Q3). This platform aims to improve the quality of university theses. During the development, a usability survey was conducted with students, achieving an average rating of 8.12 out of 10, which allowed for identifying and correcting details for improvement. One of our work’s contributions is using current web technologies combined with pre-trained BERT models, which will enable quick and easy user analysis.