Generation of Thai Spelling Distributions for Reading through the improvement of Natural Language Processing Techniques

Thai spelling distributions for reading are fundamental to understanding word components and the principles of phoneme synthesis, which are essential for accurate pronunciation and reading comprehension. This research focuses on generating Thai spelling distributions, addressing the unique challenges posed by the complexity of Thai words. These challenges include the varying number of syllables, the intricate relationships among syllable components, and the non-linear arrangement of consonants and vowels. This research proposes a methodology for generating Thai spelling distributions for reading using improved natural language processing (NLP) techniques. The approach integrates a customized dictionary of Thai consonant and vowel sequences with statistical calculations based on the Normalized Levenshtein Distance (NLD). The system evaluates a dataset of authentic Thai words sourced from the Grade 1 Basic Vocabulary List provided by the Office of the Basic Education Commission. The proposed methodology demonstrated high accuracy, closely matching human-generated spelling distribution. This research also contributes a method for generating Thai spelling distribution and a novel framework for decomposing Thai word components. This framework holds significant potential for broader applications in Thai NLP tasks. It can support children with spelling difficulties, such as those with dyslexia, by creating educational tools and customized learning programs. These resources can improve spelling and reading skills for children with special needs, making language learning more inclusive and effective.