Evidence record · UDINUS · SDGs
- Unit
- Rektorat UDINUS
- Period
- 2014
- Goals
- Record
- 30018dda-b87b-4d10-854a-1fb4c3a32b65
- Published
- 2026-10-30
Universitas Dian Nuswantoro conducted research on a technique for accessing linked open government data through a SPARQL Endpoint. The activity involved UDINUS researchers Shidik, Guruh Fajar and Ashari, Ahmad. The research focused on accessing linked open government data and producing historical time-series data concerning forest fire. The data were then used automatically as background knowledge for predicting the number of forest fire and the size of burn area through machine learning.
The source describes linked open government data as an online source of background knowledge that can provide time-series data for predicting the trend of fire disaster. The research therefore covered data access, the production of historical time-series data, the use of those data in prediction, and machine-learning processing. The prediction concerned the number of forest fire and the size of burn area. The source does not provide the dataset details, the study location, the prediction results, or numerical performance values.
Prediction performance was evaluated using mean square error and root mean square error. The research also compared Linear Regression, Neural Network and SVM algorithms using different window size. These comparisons formed part of the described evaluation of the prediction approach. The source identifies the research topic, access technique, data use, prediction subjects, evaluation measures and algorithms, but does not state which algorithm performed best or provide the results of the comparison. Universitas Dian Nuswantoro presents this activity as research involving linked open government data, historical time-series data and machine learning for forest fire prediction. The supplied RAW_GOALS value is not treated as a source fact.