Linked Open Government Data as Background Knowledge in Predicting Forest Fire

Nowadays with linked open data, we can access numerous data over the world that more easily and semantically. This research focus on technique for accessing linked open government data LOGD from SPARQL Endpoint for resulting time series historical of Forest Fire data. Moreover, the data will automatically uses as background knowledge for predicting the number of forest fire and size of burn area with machine learning. By using this technique, LOGD could be used as an online background knowledge that provide time series data for predicting trend of fire disaster. In evaluation, mean square error MSE and root mean square error RMSE are used to evaluate the performance of prediction in this research. We also compare several algorithm such as Linear Regression, Neural Network and SVM in different window size.