Redundancy reduction in environmental data set by means of an unsupervised neural networks

The acquisition of environmental data, like pollution and/or meteorological data requires the processing of a huge amount of heterogeneous data from external fields. As the number of monitoring points grows, we need a strategy to validate the acquired data and to efficiently utilize the transmission resources. An efficient way to obtain the validation-compression of the data sets is the adoption of a restricted set of samples (templates) that describe, with an assigned accuracy the whole data set. The aim of the work is to propose a validation-compression technique based on features, extracted by means of an unsupervised neural network. The paper reports the results obtained utilizing the above procedure to a real data set of a chemical pollutant. It is shown that the validation process allows a correct identification of corrupted and/or anomalous data, comparable with the human validation. Moreover the process allows a considerable reduction of transmitted data as the compression process profits the local processing of redundant data.