Improving stemming techniques for non-formal indonesian sentences using incorbiz

Computer technology emulates the human behaviour in presence of Artificial Intelligence (AI) and Natural Language Processing (NLP). The methods make computers act like humans to provide automatic answers according to the humans needs. However, the available tools for the Indonesian language in NLP are very limited, so the linguistic processing of Indonesian on computer is difficult. Linguistic computation requires a process, so the sentences can be analyzed using certain methods. The process of data pre-processing consists of case folding, filtering, tokenizing, and stemming. Stemming is one of the potential problems. The development of NLP in Indonesia is provided by stemming techniques named “Sastrawi”. The packages can be running well because of stemming on formal Indonesian sentences. On the other hand, most of the Indonesian people are using non-formal sentences in online business conversations like “sdh saya tf tlg segera proses” (I have already transferred, please proceed immediately). This research aims to improve stemming techniques for non-formal Indonesian sentences that have not been done by previous researchers. The results showed Incorbiz can find root words that have not been found by other stemming techniques. ICIC International © 2021.