BUSINESS ANALYTICS ON ACTUAL ONLINE PURCHASE: AN APPROACH USING SENTIMENT ANALYSIS BASED ON ONLINE REVIEWS

Author Name: 1. Mafas Raheem, 2. Jian Wei Cheong

Volume/Issue: 04/07

Country: Malaysia

DOI NO.: 08.2020-25662434 DOI Link: https://doi-ds.org/doilink/12.2023-43997131/UIJIR

Affiliation:

1. School of Computing, Asia Pacific University of Technology and Innovation, Kuala Lumpur, Malaysia.
2. School of Computing, Asia Pacific University of Technology and Innovation, Kuala Lumpur, Malaysia.

ABSTRACT

This paper investigates the relationship between online reviews and online purchase intention, with a focus on the electronic products industry in America. While regression and other analytics are commonly used for such analyses, there is a scarcity of research on the impact of online reviews on actual online purchases in this specific industry. The study uses KDD methodology and Tableau for insights, with Python handling data pre-processing and building a sentiment predictive model using both unsupervised and supervised techniques. The findings highlight that review quantity significantly influences actual online purchases compared to other online review elements. Bernoulli Naïve Bayes model performed well in predicting the online review sentiments towards the actual purchase (62.54%). Logistic Regression well classified the sentiment of online reviews polarity with minor errors (85.46%). Positive online reviews greatly impact Amazon purchases, especially for electronics. Sentiment analysis is crucial for businesses to grasp market trends and consumer needs, guiding effective marketing strategies. Utilizing the HEARD technique enhances sentiment analysis, boosting brand awareness, popularity, and online business revenue.

Key words: Online Reviews; Actual Online Purchase; Purchase Intention; Sentiment Analysis; Big Data; Machine Learning.