Abstract
Temperature and rainfall forecasting is one of the concerns in the agricultural sector to support people in planning suitable crops. Several techniques have been proposed previously for temperature and rainfall forecasting based on statistical analysis, machine learning, and deep learning. In our study, we propose a method using Multivariate Long-Short Term Memory (MLSTM) model to forecast monthly temperature and rainfall. The model parameters are tuned to suit the proposed problem. The model is evaluated with RMSE and MAE error measures. In addition, other forecasting models such as LSTM, MLP, and SVR are also tested to compare the effectiveness of the proposed model. Experimental results on the monthly average temperature and rainfall data set in Vietnam from 1901 to 2015 show that the MLSTM model is quite effective with the RMSE error on the temperature data set being 1.311 and MAE being 1.051, on the rainfall data set being 2.299 and 2.450.