Using neural networks for the detection of textual patterns in academic texts: A case study at the university of Nariño

Dealing manually with large volumes of textual information leads to problems in qualitative research, such as the time spent reading and processing the data and the possibility of losing critical details. In education sciences, investigations are commonly based on the qualitative paradigm, requiring substantial amounts of data to be processed textually. In this situation, an alternative for processing text to detect patterns using neural networks in the computational linguistics scenario is proposed. In this regard, qualitative research using a quasi-experimental method is performed. The quasi-experiment is done within the Qualitative Information Analysis course of the master's degree in Higher Education Teaching in 2022 with Class XIV; in such a course, six students evaluated the computational solution using material related to their working master’s thesis. The results from the application test are satisfactory due to the positive user acceptance, providing a plausible alternative to commercial computational solutions on the market.