Conceptual modeling of forest fires flashes by DataMining ontology-based. Part 2
Abstract
The heliocentric hypothesis of causes of forest fires outbreaks lias been tested. We found evidence of the correlation between the sudden arrival of charged particles from the sun and the occurrence of forest fires with a delay of one to four days. In this research, the comparative analysis was made between ANFIS and Neuron Networks in the task of searching a functional dependence between the occurrence of forest fires and the factors which characterize the solar activity. For this purpose we used several methods (R/S analysis, Hurst index, DataMining) for establishing potential links between the influx of some parameters from the sun and the occurrence of forest fires with the lag of several days. We found an evidence for a connection and developed a forecasting scenario based on the ANFIS and Neuron Network techniques. This scenario, in some cases, alliws to predict occurrences of forest fires with up to 93% accuracy.References
D’yakonov V.P. MatLab 6.5 SP1 (7.0). Simulink 5/6. Osnovy primeneniya, Izd-vo SOLON-press, 2005. — 800 s.
Abraham A. Adaptation of Fuzzy Inference System Using Neural Learning. In Fuzzy Systems Engineering: Theory and Practice (eds. Nedjah, N. & de Macedo Mourelle, L.). Springer Verlag, 2005. — P. 53–83.
Mohandes M., Rehman S. Estimation of wind speed profile using adaptive neurofuzzy inference system (ANFIS) // Appl. Energy. — 2011. — 88(11). — P. 4024−4032.
Vyklyuk Y. Simulation of spatial form of urban systems by diffusion methods // Journal of the Geographical Institute "Jovan Cvijic" SASA. — 2013. — 63, № 1. — P. 89–100. — № 2. — Р. 67–77.
Yaroslav Vyklyuk, Darko Vuković, Ana Jovanović. Forex predicton with neural network: usd/eur currency pair // Actual Problems of Economics. — 2013. — № 10. — P. 261−273.
Labitzke K. The global signal of the 11-year sunspot cycle in the atmosphere: When do we need the QBQ? Meteorolog. Zeitshrift. — 2003. — 12(4). — P. 209−216.
