Bayesian probabilistic modeling applied to environmental data, using PyMC library
Construction of a Bayesian multivariate model for prediction of Dissolved Oxygen concentration from the Piracicaba River and Paraiba do Sul River (São Paulo State - Brazil) and a Bayesian multivariate model for prediction of Benzene concentration in air polution.
The author is very grateful to the data provided by CETESB, and to Ing. Ph.D. Saverio De Vito (ENEA – Italy) for the data available.
Note about the data used in the models:
-The input data of all the models were autoscaled in the CSV spreadsheets, as a pre-treatment.
-The CSV files were divided into calibration and prediction sets, according to the number of samples.