Weather Prediction Using the Hybrid Fuzzy Logic and Long Short Term Memory (LSTM) Method Approach

Tobacco cultivation in West Lombok Regency holds an important position as an agricultural sector that makes a significant contribution to the local economy. However, this sector faces major challenges due to erratic weather patterns and the impact of global climate change. This research attempts to design a weather prediction model by adopting a hybrid methodology that combines fuzzy logic and Long Short-Term Memory (LSTM) with the aim of supporting tobacco cultivation. Weather data from the Meteorology, Climatology and Geophysics Agency (BMKG) for the West Nusa Tenggara region from January 2000 to June 2024 was used for the purposes of this research. By utilizing fuzzy algorithms, the goal is to overcome elements of uncertainty and produce initial estimates, while LSTM exploits long-term interdependencies in the data to produce more accurate predictions. The findings of this study show that combining these two approaches produces more precise and relevant weather information for tobacco farmers, resulting in increased productivity and reduced losses due to unpredictable weather conditions. This research not only contributes to increasing the productivity of tobacco cultivation, but also enriches existing scientific literature regarding the application of smart technology in the agricultural sector.