Sentiment analysis on national cultural tourism using Linear Support Vector Machine (LSVM) / Nur Haida Hanna Samsuddin
Nowadays, sentiment analysis plays a big role for many industries especially it is something related with feedback or reviews from people in cyberspace. People reviewed some products, places and others by expressing their opinion or emotion into sentences. This leads to the problem of understanding...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | https://ir.uitm.edu.my/id/eprint/55332/1/55332.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Nowadays, sentiment analysis plays a big role for many industries especially it is something related with feedback or reviews from people in cyberspace. People reviewed some products, places and others by expressing their opinion or emotion into sentences. This leads to the problem of understanding the meaning behind the texts and difficult to discover the sentiment polarity of certain words. Cultural tourism in Malaysia is lacking in terms of their promotional activities and any related authorities in the tourism industry often overlooked the reviews from tourists about cultural heritage destinations. Moreover, negative reviews may impact the national tourism. This study will perform sentiment analysis on national cultural tourism of tourists reviews on TripAdvisor website. The study will identify sentiment analysis tasks based on classification model. A classifier will be designed and developed which is Linear Support Vector Machine (LSVM). Lastly, the accuracy of the proposed classifier will be tested. Therefore, the chosen technique is classification and the algorithm that will be applied in the classification process is Linear Support Vector Machines (LSVM). The output will be the accuracy of the LSVM model and the visualization of sentiment analysis of new data that user will choose in the prototype. The accuracy achieved from the project is 80%. The classifier is claimed to be bad classifier because AUC-ROC gained from the experiment is 0.5. In future, it is recommended to experiment with different algorithm or kernel of Support Vector Machine. The volume of data should be large as it can generate better result of classification method. |
|---|