Leveraging Jaro-Winkler for Enhanced Nazief-Adriani Banjar Text Stemming

The Banjar language, predominantly spoken in South Kalimantan, Indonesia, poses significant challenges in natural language processing (NLP) due to its complex morphological and phonological features. This study aims to optimize the Nazief-Adriani stemming algorithm to more effectively process Banjar language texts. The original algorithm, tailored for the Indonesian language, often results in over stemming and under stemming when applied to Banjar words. To address these issues, this study integrates the Jaro-Winkler method for string similarity measurement and implements specific Banjar morphological rules into the stemming process. This study utilizes language analysis tools such as Python’s Natural Language Toolkit (NLTK) and NumPy for data processing and analysis. This study’s results show that the optimized algorithm achieves a stemming accuracy of 79,34%, indicating a significant improvement over the original method. This research contributes to the development of more accurate NLP tools for regional languages like Banjar and suggests further exploration of morphological variations to enhance algorithm performance.