Abstract:
The increasingly widespread use of the internet allows companies to gain insights regarding product development. Online
reviews can provide an overview of how customers need a product, and this can help companies to understand what customers
like and dislike about their products. This paper proposes a methodology for identifying consumer needs by analysing online
reviews using the lexicon-based method and topic modeling with Non-negative Matrix Factorization (NMF). The main idea of this
paper is to translate the results of aspect-based sentiment analysis into consumer needs with its priority ranking. The methodology
is applied in an online review of a laptop product. Based on the study results, it is shown that the observed features based on
customer reviews are battery, storage, screen, price, performance, keyboard, and design, with each topic having overall positive
sentiment. Based on the topics and their sentiment, the prioritized needs are good screen strength, a fast and responsive processor,
and responsive keyboard keys.