Otomasi Klasifikasi Taksonomi Pengetahuan pada Knowledge Management System (KMS) Menggunakan Algoritma Natural Language Processing (NLP)

Managing knowledge assets within a Knowledge Management System (KMS) often faces constraints during the manual document classification stage. Manual processes are not only time-consuming but also prone to subjectivity and inconsistent taxonomic labeling. This research aims to automate knowledge taxonomy classification in KMS by integrating Natural Language Processing (NLP) techniques. The methodology includes text preprocessing stages (such as case folding, tokenizing, filtering, and Sastrawi stemming) and feature extraction using Term Frequency-Inverse Document Frequency (TF-IDF). Test results on 100 Indonesian text documents indicate that the application of AI through classification algorithms achieves an average accuracy rate of 85%. Furthermore, efficiency analysis shows a significant reduction in document processing time, from an average of 120-300 seconds to less than 0.5 seconds per document. The stemming stage proved crucial as it increased system accuracy by 23% through a 72.2% reduction in word features. This study concludes that NLP integration effectively enhances the scalability and accuracy of organizational knowledge management while minimizing administrative workload.