Sentiment Analysis of Airline Reviews Using Naive Bayes Algorithm / Zaiton Mohd Napiah
Reviews and customer’s feedback play a vital role for companies to monitor their brand, improve customer service or gain customer's trust. Even though technologies have evolved over the years, some companies such as airline companies still use traditional methods to collect reviews by using sur...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2021
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/54673/1/54673.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Reviews and customer’s feedback play a vital role for companies to monitor their brand, improve customer service or gain customer's trust. Even though technologies have evolved over the years, some companies such as airline companies still use traditional methods to collect reviews by using surveys which may lead to human error. This includes bias and random response. Furthermore, with the advancement of technology, people tend to express their satisfaction or dissatisfaction through social media such as Facebook or Twitter which is becoming more popular even among the older generations. Therefore, this project aims to help the company to classify and visualize the sentiments of the reviews by collecting the reviews from Twitter or uploading a file containing reviews to the system. The sentiment of the reviews will be classified into two categories which are positive and negative. The overview for the methodology is divided into three phases which are preliminary phase, design and implementation phase and evaluation phase. Several classifier models have been built by using Naive Bayes algorithm during the design and implementation phase where the model that has the highest accuracy has been chosen for this project. On top of that, functionality testing has been done on the system to ensure that there are no problems especially in the visualization component of the system. Some recommendations have also been given to improve the project such as training the model with other languages such as Malay and analyse reviews from other social media such as Facebook, hi conclusion, this project may help to ease the burden of airline companies as they no longer need to analyse the reviews one by one. |
|---|