Predictive analytics for the sentiment of malaysian place of interest using machine learning models
Sentiment analysis is a method of automatically identifying sentiments expressed in online interactions with the aim of evaluating users or customers' opinions on a product, brand, or service. It helps companies gain valuable insights and respond to their customers more efficiently. As people u...
Saved in:
| Main Author: | |
|---|---|
| Format: | Undergraduates Project Papers |
| Language: | en |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://umpir.ump.edu.my/id/eprint/46447/1/Predictive%20analytics%20for%20the%20sentiment%20of%20malaysian%20place%20of%20interest%20using%20machine%20learning%20models.pdf https://umpir.ump.edu.my/id/eprint/46447/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Sentiment analysis is a method of automatically identifying sentiments expressed in online interactions with the aim of evaluating users or customers' opinions on a product, brand, or service. It helps companies gain valuable insights and respond to their customers more efficiently. As people use social media, forums, blogs, and the web to express their opinions on various discussion topics, these channels have become an ideal domain for utilizing customer sentiment analysis. The focus of this study is to conduct Natural Language Processing (NLP) on tweets and make a better classification of sentiment using Malaya. Furthermore, this study also trains three machine learning algorithms to predict the sentiment of textual data. Moreover, this study also creates dashboard to visualize social media insights and suggest recommendations based on the insights. The study gathered users or customers feedback from Twitter on 1st January 2023 to 1st March 2023 using the social media monitoring software, Determ, which retrieves tweets in real-time based on search terms, time, users, and likes. The tweets containing feedbacks and responses were organized into tables and saved as a CSV file. Subsequently, the study proceeded with the pre-processing stage to handle missing and erroneous values in the data. Additionally, several Natural Language Processing (NLP) techniques were employed to pre-process the text data, as part of the machine learning process. The data was then divided into training and testing sets, and was trained using three different supervised learning algorithms, namely Support Vector Machine, Random Forest, and Naive Bayes. Finally, the performance of prediction of each model was compared to identify the most accurate one, and based on the analysis, it was concluded that Support Vector Machine exhibited the best performance in terms of accuracy, recall score, F1 score, and precision score. Furthermore, it is worth noting that this sentiment analysis research is extended to analyze sentiments expressed in texts written in Malay language by utilizing the Natural Language-Toolkit library for Bahasa Malaysia, powered by Tensorflow and PyTorch. Regarding the outcomes of the customer sentiment analysis, there are some recommendations that can be adopted to enhance the effectiveness of the study. For instance, the analysis can be extended to include customer feedback data collected from social media platforms such as Facebook, Instagram, Tik Tok, web and forums. Additionally, the performance of the customer sentiment analysis model can be improved by leveraging deep learning techniques. |
|---|