ANALYSIS OF WEATHER CHANGES FOR ESTIMATION OF SHALLOT CROPS FLUCTUATION USING HIDDEN MARKOV

Abstract

Climate change has an impact on increasing the temperature of the earth's surface or what is known as global warming. The impact of global warming will affect the pattern of precipitation, evaporation, water run-off, soil moisture and climate variations which are very volatile can threaten the success of horticultural production, especially shallots. Shallots are a strategic commodity but are strongly influenced by fluctuations in production. The development of shallots is one of them constrained by the weather/climate which affects the production of shallots. From these constraints, shallots are also a commodity that contribute significantly to inflation. Hidden Markov Models (HMM) is one of the stochastic processes when the future only depends on condition now, in markov chain all of the element observable, and the probability move to another probability. Prediction and estimation of shallot crops with rainfall input, temperature, and humidity is done with data starting in 2016 until 2020. Estimated shallot crops follows the optimum movement pattern of prediction shallot in each of each variable. The planting months that are usually carried out in the two districts are around February, May, June and September the lowest shallot crops in April or May because transition of rainy to dry season. And the highest shallot crops in October or November. The best accuracy of estimation is rainfall factor with MAPE 5,89% with high accuracy category while 5,84% in MAPE temperature and in 5,55% in humidity factor in category high.