Text Mining Techniques and Pre-trained Models Applied to the Issues in Pictograms: A Systematic Literature Review

Currently, a fundamental action for people of any age is proper communication with their environment, so new ways are sought to perform them, such as using Augmentative and Alternative Communication (AAC) for people with Complex Communication Needs (CCN). This type of communication through technological tools allows people with CCN to transmit and receive messages through pictograms. However, these people have issues with pictogram selection in assisted communication systems, such as the low quality of sentences produced, even though some algorithms and methodologies allow pictograms to be found in a reasonable time. For this reason, this article presents a Systematic Literature Review (SLR) to synthesize primary studies and identify the data science tools, methods, and approaches used in pictograms to facilitate communication in people with Complex Communication Needs. 508 articles were extracted, of which only 42 were selected after applying the exclusion and inclusion criteria and the quality assessment. The results obtained could help establish research gaps and thus contribute to creating new techniques and algorithms in data science based on the above topics.