Abstract – Suitable soil temperature predictions can help the farmers and producers in providing valuable information for deciding the right time for crop cultivation and harvesting. Due to non-linearity in climatic physics, neural networks are suitable to predict these meteorological processes. Back Propagation is the most important algorithm to train a neural network. It is a systematic gradient descent method that suffers from local minima problem, scaling problem, long training times, etc. In this article, an efficient soil temperature prediction model is proposed using neural network based genetic algorithm technique to solve these problems. The results are very encouraging.
