Makassar, Indonesia, August 2nd, 2022/asds/. In times of disaster, information can save lives. But today, information does not only come from official reports—it flows rapidly through social media. Every tweet, post, or message reflects how people feel, what they need, and how they respond to crises. The challenge is turning this flood of data into meaningful insights. This is where artificial intelligence (AI) steps in, offering new ways to support disaster management and strengthen collaboration, in line with SDG 17: Partnerships for the Goals.
A recent study titled*“Comparison of Support Vector Machine Based on Fasttext Without and With Firefly Optimization Parameters for Disaster Sentiment Analysis in Indonesia”* explores how AI can analyze public sentiment during disasters.
Indonesia, located on major tectonic plates, is highly prone to natural disasters such as earthquakes, floods, and volcanic eruptions. During such events, people often use platforms like Twitter to share experiences, ask for help, or express emotions. These digital footprints become valuable data for understanding public needs in real time.
The study collected more than 1,600 tweets related to disasters in Indonesia and used machine learning to classify them into positive or negative sentiments. Positive sentiments often included support, prayers, and solidarity, while negative ones reflected fear, complaints, or distress. This classification helps authorities quickly identify which areas need urgent attention.
To process the data, researchers used a technique called Support Vector Machine (SVM)—a machine learning method that can separate data into categories. However, SVM alone has limitations because it requires precise parameter settings to perform well. To overcome this, the researchers applied Firefly Optimization, an algorithm inspired by the way fireflies attract each other using light.
The results were remarkable. Without optimization, the model achieved only 61.3% accuracy, meaning many messages were misclassified. But after applying Firefly Optimization, accuracy increased dramatically to 89.1%. This improvement shows how combining different technologies can significantly enhance performance
So why is this important? Because accurate sentiment analysis allows governments, NGOs, and disaster response teams to work more effectively together. For example, if many tweets express urgent needs for food or shelter, responders can act faster. If public sentiment shows confusion or misinformation, communication strategies can be improved.
This is where SDG 17: Partnerships for the Goals comes into play. Disaster management requires cooperation between multiple stakeholders—government agencies, researchers, technology developers, and communities. AI tools like this study demonstrate how technology can bridge these actors, enabling them to share data, insights, and resources more efficiently.
Moreover, this approach highlights the power of data-driven decision-making. Instead of relying only on traditional reports, policymakers can use real-time insights from social media to understand the situation on the ground. This leads to faster, more targeted, and more inclusive responses.
At a broader level, this study also shows how local research can contribute to global challenges. By developing solutions tailored to Indonesia’s context, researchers create models that can be adapted to other disaster-prone regions around the world.
In conclusion, analyzing social media with AI is more than a technological innovation—it is a tool for saving lives and strengthening collaboration. By transforming scattered online voices into actionable knowledge, we move closer to achieving SDG 17: Partnerships for the Goals, where collective efforts and shared technologies help build more resilient societies.
DOI: https://doi.org/10.30598/barekengvol18iss3pp1791-1802
Contact: Prof. Sri Astuti Thamrin
+6285298934579




