Big Data Analytics for Predicting Coral Bleaching in Samui Island Area, Suratthani Province

Abstract

This research investigates coral bleaching by collecting sea temperature data and seawater acidity. Through the technology of the Internet of Things together with LoRa and satellite data acquisition. We test the effectiveness of predicting coral bleaching in the Samui island area of Suratthani province which consists of an SVM, Naive Bayes, Logistic regression model, and data visualization using spatial analysis techniques. The result of this study was found that three important parameters for the development of LoRa devices are spreading factor, bandwidth, and code rate. The most important parameter to set up and affect the RSSI value is the spreading factor. After we studied and tested the system with LoRa, we tested the effectiveness of the model in various datasets. We found that the SVM model had accurate on data from a pontoon. The model's performance was tested using two techniques: split test, and K fold crossvalidation. We visualize of information gathered from various sources shows the risk of coral bleaching during the data collection period. The most clearly shown source is the pontoon source.