Abstract
Online reviews play a significant role in the success of a business. Deep learning models have emerged as crucial tools in this domain, with one-dimensional Convolutional Neural Network (1D CNN) being commonly used. However, this paper proposes a novel approach utilizing a Two-Dimensional Convolutional Neural Network (Att + 2D CNN) with attention mechanism, which effectively captures the dimensionality of the input text, resembling a 2D matrix. To further enhance the model’s performance, we employ pretrained word embeddings, specifically GloVe and Word2Vec. We thoroughly analyze the performance of these embeddings in conjunction with deep learning models. Remarkably, our proposed method, leveraging 2D CNN with attention, consistently achieves superior accuracy when compared to other models, specifically on Amazon Cell Phone reviews and Amazon Kindle reviews datasets, for both balanced and unbalanced natures. By employing this novel methodology, we demonstrate the ability to extract valuable insights from online reviews, enabling businesses to make informed decisions.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Correction Statement
This article has been corrected with minor changes. These changes do not impact the academic content of the article.
Additional information
Notes on contributors
Amrithkala M. Shetty
Amrithkala M. Shetty is a research scholoar in the department of computer science, mangalore University. Area of expertise include Sentiment analysis, Depp learning and machine learning.
Mohammed Fadhel Aljunid
Mohammed Fadhel Aljunid Data Scientist and Systems Developer for the Investment Sector at the Higher Agricultural and Fisheries Committee, in Yemen. an accomplished professional in the field of Computer Science, with a special focus on Data Science, Recommendation Systems, Natural Language Processing (NLP), Sentiment Analysis, Machine Learning and Perception, and AI.
D. H Manjaiah
D. H Manjaiah working as a senior professor in the department of computer science, Mangalore University. Area of specialization includes Network, Routing, Ad Hoc Networks, Wireless Networks, Routing Protocols, Computer Networking, Network Communication, QoS, Recommender systems and NLP.