Assessment on Tubalan Marine Sanctuary in Conserving Existing Coastal Habitat Using Data Mining Techniques

Abstract

This paper predicts the accuracy of the marine coastal ecosystem health evaluation dataset based on the conducted assessment by the Department of Environment and Natural Resources (DENR). The study conducted is within a marine sanctuary in Malita, Davao Occidental, where fishing is the means of livelihood. The use of data mining techniques can help in the close monitoring of the reefs in the area. The marine coastal ecosystem health evaluation dataset used in this study consisted of seven variables, with 966 instances and were assessed using three data mining algorithms, namely the Naive Bayes, K-Nearest Neighbor (KNN), and C4.5 algorithms. The results of the study show that the Naïve Bayes, KNN, and C4.5 algorithms obtained 91.24%, 98.05%, and 96.86% prediction accuracies, respectively, where the identified optimal algorithm for prediction is the KNN algorithm. Finally, this study paved the way to determine indicators of a healthy marine ecosystem through mining targets or high economic value fishes.