Modeling of Environmental Factors for Finding Optimal Conditions on Cultivating Farm Products

Since the number of farmers has been decreasing recently, shortage of the labor force is a serious problem in many farmhouses. In order to solve this problem, it is necessary to realize the system to support farmer´s works in low costs. The purpose of our research is to construct the system which can predict the farmland environment in the near future. In this research, we focus on the control of soil wetness and temperature. We formalize a model for expressing the rule for predicting temperature and soil wetness from the latest environmental data of farmhouse. We show that the rule can be generated by the machine learning algorithm ID3. We research the confidence of each prediction by comparing data obtained from the experiment of cultivating farm products using a greenhouse. Based on the result, we research for finding environmental factors which are needed to create the hypothesis for the prediction of the environment transformation.