CRC – Clothing Review Classification using sentiment analysis / Nur Suhailayani Suhaimi ... [et al.]
Safehse is a clothing brand established in 2020. Safehse clothing highlights the essence of South Korean streetwear and puts it under the spotlight. The customer text review is one of the features that Safehse have to play a role in helping the customer to make their purchasing decision. However, Sa...
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/93560/1/93560.pdf https://ir.uitm.edu.my/id/eprint/93560/ https://jamcsiix.uitm.edu.my/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Safehse is a clothing brand established in 2020. Safehse clothing highlights the essence of South Korean streetwear and puts it under the spotlight. The customer text review is one of the features that Safehse have to play a role in helping the customer to make their purchasing decision. However, Safehse encountered a few problems with the current process of identifying genuine or fake reviews, time-consuming to classify positive and negative from the customer text review, and misleading or misunderstanding reviews that need to be clarified for the customers. In order to reduce and minimize the problem, Safehse needs to use the classification of text reviews by using sentiment analysis. This research project aims to develop a system to classify text reviews for Safehse, which can identify positive and negative reviews. The sentiment analysis technique used for this text review classification is supervised machine learning that anticipates occurrences by combining what it has learned from prior and current data with labels. The outcome of this text review classification is to display the categorized reviews with the calculated tokenization. As for the result of this project, the system will display the categorized review with the classification of positive and negative. The project will also display the genuine or fake review categories with the percentage of criteria from the data training. In average, more 75 % of sample data are correctly classified based on their pre-defined classes and more 55 % of data precisely categorized into fake or genuine label. For future enhancement of this project, a mobile application feature can be added to ease the text classification process quickly. |
|---|