Computational Parsing of Morphological Patterns in Digital Philology Courseware
Abstract
The goal of this research is to provide a new computational framework for analyzing morphological patterns, designed for use in digital philology courseware. There is a computer framework called MorphoScribe, an accessible computer program that utilizes deep learning to identify patterns and segment data based on predefined rules. Using Universal Dependencies (UD) Treebanks makes this possible. MorphoScribe is the parts that make it possible. The software was tested on UD datasets with ten different languages, achieving an average morphological parsing accuracy of 94.2%. The testing that was done made this possible. Another thing to consider is that its precision and recall rates were higher than 93% and 92%, respectively, compared to other products. When it came to the error rates for morpheme boundary recognition, the system was able to lower them by 37% compared to the baseline models. According to the results of educational trials with 120 pupils, parsing activities were finished 32% faster, and morphological analysis abilities were 42% better. It was clear that both changes were for the better. Ninety-five percent of the students who took MorphoScribe's interactive courses reported being satisfied with the platform, as indicated by their responses. The findings presented in this research demonstrate that MorphoScribe not only enhances morphological parsing but also improves the learning experience in digital philology courseware. This is demonstrated by the fact that MorphoScribe helps children learn more effectively.