Classification and visualization of E-commerce product reviews comparison using support vector machine / Nuwairah Aimi Ahmad Kushairi, Khyrina Airin Fariza Abu Samah and Nurul Atirah Ahmad
E-commerce is significantly growing as a platform for online shopping, offering convenience and costsaving benefits. Especially in Malaysia, Shopee is one of the leading e-commerce platforms. These days, online product reviews play a crucial role in influencing consumer behaviour by building trust,...
Saved in:
| Main Authors: | , , |
|---|---|
| Format: | Book Section |
| Language: | English |
| Published: |
Faculty of Computer and Mathematical Sciences
2023
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/94178/1/94178.pdf https://ir.uitm.edu.my/id/eprint/94178/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | E-commerce is significantly growing as a platform for online shopping, offering convenience and costsaving benefits. Especially in Malaysia, Shopee is one of the leading e-commerce platforms. These days, online product reviews play a crucial role in influencing consumer behaviour by building trust, identifying customer needs, and improving satisfaction. 181 out of a 186 respondents questionnaire survey agreed that they rely on product reviews before purchasing any product. Nevertheless, 179 respondents agreed that not all product reviews are helpful when shopping online. It becomes time-consuming to read through the reviews, especially when it is not product related. Moreover, an abundance of reviews can lead to information overload, which exhausts customers to decide. Therefore, this study aims to classify the comparison of useful and not useful product reviews from Shopee using Support Vector Machine (SVM) and visualize the comparison. Users can enter up to six product links, and the system will classify reviews based on review text, star rating, duplicated spam, and sentiment score. Testing showed 96.8% accuracy and passed all functionality test cases. Mann-Whitney U Test obtained a p-value of 0.008, indicating a significant difference in evaluation time over manual evaluation, proving its potential in aiding purchase decisions. |
|---|