The carbon cycle is an important component of life on Earth, and the decomposition of compounds is part of the carbon cycle. One key component of this part of the process is the decomposition of plant material and woody fibers. The purpose of this report is to establish a fungal decomposition rate prediction model to evaluate the impact of environmental changes on fungal activity, and therefore on the ecosystem. This paper aims to build two models: the model :Fungi Decomposition Prediction Model Based on BP Neural Network Algorithm; the model :Colony Evolution Model Based on Time Series Algorithm. In addition, the present report discusses the impact of colony diversity on ecosystems. Different microbial community construction has different effects on the decomposition, thus it can promote the decomposition of litters to a certain extent. And the diversity of the fungal community is conducive to sustainable development of the ecological environment.
Keywords: ARMA; BP Neural Network; Stationary test; Time Series.