Exploring Japanese Language Reviews on Amazon Products: Leveraging Word Embedding and Deep Learning Methods

Amidst growing global connectivity, businesses started operating across borders. Most people rely on translators to communicate and understand the languages. Japan is one of the major economic players and also a multifaceted language with unique characters and grammar structures that differ significantly from other Western languages. Entrepreneurs struggle to comprehend their goods’ performance and seek feedback for improvements. Hence, the translator aids in bridging the gap by accurately conveying the meaning between Japanese and other languages, thereby facilitating communication with Japanese clients or customers. This work uses Google Translate to rephrase the Japanese Amazon Product Reviews to English. Entrepreneurs can use these reviews and ratings to improve their products by blending Word Embedding and Deep Learning (DL) techniques. The word embedding techniques like Word2Vec and FastText for feature extraction and DL models like LSTM(Long Short-Term Memory) and Bi-LSTM(Bidirectional-LSTM) are used for the classification of product reviews. According to the findings, the blend of Word2vec and LSTM outperformed all the other models with an accuracy of 97.13%.