Optimizing Extreme Learning Machine for Drought Forecasting: Water Cycle vs. Bacterial Foraging

Machine learning (ML) methods have shown noteworthy skill in recognizing environmental patterns. However, presence of weather noise associated with the chaotic characteristics water cycle components restricts capability standalone ML models modeling extreme climate events such as droughts. To tackle problem, this article suggests two novel hybrid based on combination machine (ELM) algorithm (WCA) and bacterial foraging optimization (BFO). The new models, respectively called ELM-WCA ELM-BFO, were applied to forecast standardized precipitation evapotranspiration index (SPEI) at Beypazari Nallihan meteorological stations Ankara province (Turkey). performance proposed was compared those ELM considering root mean square error (RMSE), Nash-Sutcliffe efficiency (NSE), graphical plots. forecasting results for three- six-month accumulation periods showed that is superior its counterparts. NSE SPEI-3 testing period proved improved drought accuracy up 72% 85% stations, respectively. Regarding SPEI-6 results, achieved highest RMSE reduction percentage about 63% 56%