BenSW: A Standard Dataset for Bengali Stop Word Detection

In the Bangla Language Processing system, stop word is the least contributed and resourceful domain than any other language. Stop word is used for sentence formation purposes without holding any significant value in expressing the semantic meaning of a sentence. The detection and removal of Bengali stop words is an essential preprocessing task in some fields such as sentiment analysis, text mining, information retrieval, text categorization, search index and retrieval system, text summarization. But there is no standard Bengali stop word list available, also no standard dataset is publicly available with the solution of stop word-related special case problem where sometimes stop word act as a content word contextwise and viceversa. So, in this paper, a standard dataset consists of 20k sentences that can solve the stop word-related special case problem, are introduced and made publicly available on GitHub that can motivate the researchers to contribute in this field and enrich the resource of the Bangla Language Processing system. This is the first publicly available Bengali stop word standard dataset, also the dataset is experimented with by applying different approaches and machine learning classifier algorithms, and the achieved accuracy is commendable.