Model for predicting the occurrence of soil compaction

Solving the problem of preserving and increasing soil fertility, timely determination causes its deformation require improved environmental forecasting. The development quantitative approaches in ecology is facilitated by availability data on types properties, understanding ongoing processes existing information technologies. In this article, we give an example designing training a deep neural network for ecological forecasting date occurrence compaction, as one relevant parameters further research. Using properties compacted soils, show that networks are able to predict both short-term risk compaction long-term dynamics. Against background methods, have better performance potential create integrated monitoring systems based them.