Abstract:
In today’s digitized world, reviews and blogs play an important role in getting the knowledge about the best products on the market. These reviews and blogs contain sentiments of peoples related to the products. In the field of natural language processing, sentiments are classified under the area of sentiment analysis and have always been the active area of research. The labeled data is considered as source domain and is trained for sentiment classification. The sentiment value change from domain to domain due to which when the classification is performed on cross-domain dataset, it produces the result differently. In this paper, we presented the cross-domain semantic analysis on product review dataset by capturing the semantic feature with word embedding algorithm and processing it with deep learning neural network to improve accuracy for cross-domain sentiment classification tasks.
Keywords: