Probabilistic estimation of stop words from document-term matrix for topic analysis application.
Nami Harada, Yutaka Kano · Research Square · 2024
Abstract In topic analysis, it is important to remove stop words to make inference about topics. Prior knowledge about the data is necessary to remove stop words. Stop words imply no contribution and interference of topic estimation. There is a suggestion of probabilistic inference of stop words. Probabilistic inference of stop words without prior knowledge about the data will assist in topic analysis application to unfamiliar languages and other various data fields.