Since the level of disturbance increases due to rapid population aggregation and industrial development, long-term monitoring is critically important for the sustainable management of ecosystems. Time-series data would be obtained in a great amount in the future. Consequently, the temporal networks would be mostly feasible in analyzing complex time series data often observed in ecological processes. ANNs have been reported to be efficient in prediction and classification of community and ecosystem data.
Temporal networks could be efficiently implemented to analysis of large scale data on the national and international basis. The trainable variables could be selected from various issues that are closely related with spatial and temporal development (e.g., community dynamics, food web, nutrient cycle) in a global scale. The relationships of environmental changes with the corresponding community development in temporal sequences could be further elucidated through the training by the temporal networks and the accompanying sensitivity analyses.
Along with development of computation techniques, a large amount of data could be processed as input data on the real-time basis on the internet system (see Internet). For instance, behavioral data that are continuously measured in short time intervals at various samples could be efficiently processed for on-line monitoring for risk assessment as demonstrated in RSOMs.
See also-. Hopfield Network; Internet; Multilayer Perceptron; Self-Organizing Map.
Was this article helpful?