Sentiment Analysis and Consumer Purchase Intention Prediction Based on BERT and DBLSTM

Sentiment Analysis and Consumer Purchase Intention Prediction Based on BERT and DBLSTM

Changzheng Yang (Ocean University of China, China), Mengmeng Guo (Renmin University of China, China), and Muhammad Asif (National Textile University, Pakistan)
Copyright: © 2025 |Pages: 22
DOI: 10.4018/JOEUC.372206
Article PDF Download
Open access articles are freely available for download

Abstract

With the rapid growth of e-commerce and social media, analyzing consumer sentiment and predicting purchase intentions are vital for understanding behavior and optimizing marketing strategies. This study proposes an integrated model combining Bidirectional Encoder Representation from Transformers (BERT) and Deep Bidirectional Long Short-Term Memory Network (DBLSTM). BERT efficiently extracts semantic features from consumer reviews using self-attention mechanisms, while DBLSTM captures sequential dynamics, leveraging temporal dependencies for predicting purchase intentions. The model was evaluated on real-world datasets and compared against other deep learning models. Results showed that the BERT-DBLSTM model outperformed others in sentiment analysis accuracy and purchase intention prediction, demonstrating higher generalization and prediction accuracy. This approach provides enterprises with precise market insights, enabling improved marketing strategies, enhanced user satisfaction, and increased conversion rates.
Article Preview
Top

Introduction

In today’s global business environment, e-commerce has become essential, continually advancing and expanding. With a growing preference for online shopping, consumers are increasingly turning to e-commerce to meet their needs, making it a fundamental aspect of contemporary life. Yet, even with its rapid growth, the e-commerce sector faces significant challenges, with one of the foremost being the intricate relationship between consumer trust and purchase intention (Yu et al., 2021). Simultaneously, social media have emerged as key platforms where people share their opinions and emotions. Effectively extracting consumer sentiment and gaining a deeper insight into the decision-making processes behind their purchases have become essential goals. Sentiment analysis, a technique in natural language processing (NLP), takes on an important role in achieving these objectives (Grover, 2022); it is used to extract sentiment information from text to assist companies in understanding consumers’ emotional tendencies and attitudes. Concurrently, consumer purchase intent prediction represents a valuable tool for gaining insight into the consumer decision-making process. The integration of these two approaches can facilitate the acquisition of more comprehensive market insights and the optimization of marketing strategies.

Sentiment analysis, a significant NLP technique, is becoming a crucial tool for comprehending consumer behavior and market dynamics. Sentiment analysis is a technique used to extract sentiment information from text, identify the people’s affective tendencies (e.g., positive, negative, or neutral), and categorize them. Moreover, sentiment analysis can be expanded to incorporate more advanced methods capable of detecting specific emotions, intentions, or nuanced emotional expressions (e.g., anger, happiness, or sarcasm), as well as identifying context-specific sentiments often found in areas like product reviews (Cambria et al., 2017). Sentiment analysis system architecture (Devika et al., 2016), including data collection, data cleansing, text coding, feature extraction, training models, and evaluating models, is shown in Figure 1. Basing their analysis on this overall system architecture, researchers can extract valuable information from textual content to personalize the sentiment expressed in the textual data.

Figure 1.

Working process of sentiment analysis

JOEUC.372206.f01

In recent years, deep learning has advanced considerably in both social media and e-commerce fields. It is particularly effective in analyzing consumer behavior, conducting sentiment analysis, and predicting purchase intentions (Alzahrani et al., 2022). All these methods largely rely on the powerful self-learning capabilities of deep learning networks, including feature extraction, contextual modeling, and big data processing. These capabilities facilitate more effective extraction of emotional information from text and enhance the comprehension and prediction of consumer purchase decisions. Meanwhile, these research advancements come with a range of challenges and issues. For instance, the exponential increase in data volume, along with the diversity and complexity arising from multiple data sources, will require more advanced computational resources and greater processing power to support deep learning model. Furthermore, e-commerce product reviews comprise a multitude of emotional and semantic elements, necessitating the development of sophisticated deep learning models to effectively grasp and comprehend their nuances. Additionally, e-commerce is a multidisciplinary field that requires the integration of expertise from various disciplines, providing enterprises with a broader range of solutions.

Complete Article List

Search this Journal:
Reset
Volume 37: 1 Issue (2025)
Volume 36: 1 Issue (2024)
Volume 35: 3 Issues (2023)
Volume 34: 10 Issues (2022)
Volume 33: 6 Issues (2021)
Volume 32: 4 Issues (2020)
Volume 31: 4 Issues (2019)
Volume 30: 4 Issues (2018)
Volume 29: 4 Issues (2017)
Volume 28: 4 Issues (2016)
Volume 27: 4 Issues (2015)
Volume 26: 4 Issues (2014)
Volume 25: 4 Issues (2013)
Volume 24: 4 Issues (2012)
Volume 23: 4 Issues (2011)
Volume 22: 4 Issues (2010)
Volume 21: 4 Issues (2009)
Volume 20: 4 Issues (2008)
Volume 19: 4 Issues (2007)
Volume 18: 4 Issues (2006)
Volume 17: 4 Issues (2005)
Volume 16: 4 Issues (2004)
Volume 15: 4 Issues (2003)
Volume 14: 4 Issues (2002)
Volume 13: 4 Issues (2001)
Volume 12: 4 Issues (2000)
Volume 11: 4 Issues (1999)
Volume 10: 4 Issues (1998)
Volume 9: 4 Issues (1997)
Volume 8: 4 Issues (1996)
Volume 7: 4 Issues (1995)
Volume 6: 4 Issues (1994)
Volume 5: 4 Issues (1993)
Volume 4: 4 Issues (1992)
Volume 3: 4 Issues (1991)
Volume 2: 4 Issues (1990)
Volume 1: 3 Issues (1989)
View Complete Journal Contents Listing