Public Sentiment on Awareness of Climate Change Based on Support Vector Machine

Climate change has threatened human society and natural ecosystems, yet public opinion surveys have found that public awareness and concern are very deficient. If society is unaware of climate change, activities such as open burning, deforestation, and releasing excessive carbon dioxide gases would not be reduced. There are several methods to detect public opinion on climate change, and one of the convenient and efficient methods is conducting sentiment analysis on Twitter. This study uses machine learning techniques to collect and analyze public opinion on climate change from Twitter. Due to the increasing occurrences of natural disasters worldwide, understanding public awareness of climate change is crucial. The objective of the study is to analyze public sentiment on the awareness of climate change based on the Support Vector Machine (SVM) algorithm. The methodology for the study consists of several phases: data collection, pre-processing, labeling, feature extraction and classifier evaluation. The evaluation results indicated that SVM achieved a high accuracy of 91% with an 80:20 data split. The SVM classifier model has also produced high precision, F1-score, and recall results. The government could use the study results and non-governmental organizations (NGOs) to help them spread awareness on climate change issues. Future work will improve the classifier by analyzing non-English tweets and using SentiWordNet to handle word ambiguity in the messages.