Sentiment analysis for cruises in Saudi Arabia on social media platforms using machine learning algorithms

Area of Science:

• Natural Language Processing
• Machine Learning Applications
• Social Media Analytics

Background:

• Social media platforms are crucial for community discussions and event sharing.
• Understanding public sentiment towards emerging tourism experiences, like Saudi cruises, is vital.
• Sentiment analysis (SA) provides a method to gauge public opinion from online data.

Purpose of the Study:

• To evaluate the quality of sentiment analysis for Saudi cruise impressions.
• To create and analyze datasets from Instagram, Snapchat, and Twitter for Saudi cruise feedback.
• To understand passenger and viewer opinions on their Saudi cruise experiences.

Main Methods:

• Collected and cleaned 1200 social media samples from Instagram, Snapchat, and Twitter.
• Classified data into positive or negative sentiments using machine learning algorithms: MLP, NB, RF, SVM, and voting.
• Compared algorithm performance across different datasets.

Main Results:

• The Random Forest (RF) algorithm achieved the highest classification accuracy, reaching 100% with over-sampled Snapchat data.
• All tested machine learning algorithms demonstrated high accuracy in sentiment classification.
• Analysis indicated that 80% of sentiments expressed about Saudi cruises were positive, with 20% being negative.

Conclusions:

• Sentiment analysis is effective for gauging public opinion on Saudi cruise experiences.
• The Random Forest algorithm shows superior performance for this specific sentiment analysis task.
• Public reception of Saudi cruises, based on social media, is predominantly positive.