Key points are not available for this paper at this time.
Lemmatization is a natural language processing (NLP) based text normalization technique that effectively improves data consistency and assists in context interpretation. Nevertheless, due to the highly inflected nature and morphological richness of linguistics, lemmatization in Bangla NLP holds a thorny challenge. In this research work, we take the challenge to build a comprehensive Bangla lemmatizer. We concreted homogeneous resources by compiling heterogeneous resources, assembled root words, collected linguistic rules, and applied the Trie along with "Longest Substring Search by Removing Affix (LSSRA)" to develop the lemmatizer. Our system aimed to lemmatize words based on their parts of speech within a given sentence and utilized the sequences of suffix marker occurrence according to the morpho-syntactic values by operating the longest suffix stripping methodology. The lemmatizer achieved an accuracy of 96.90% and experimented against a manually annotated test dataset.
Islam et al. (Thu,) studied this question.