Android Module for atmospheric pollutants concentration estimator for custom E-Nose array. The E-Nose produces an array of tensions, in millivolt, generated by the electrodes. Those readings will be the input for the estimators,that produce the values of pollutants

The library contains 2 estimators:

• Linear output with non linear temperature correction
• Neural network generated calibration

The library incluses a StringParser class, an utility used to parse the output of the E-nose.

A factory class is used to allocate a new estimator. The new estimator reads specific E-Noses settings from a json text file for an easy setup of new/old devices. The core of the library is based on 2 classes: LinearCalibration, NeuralCalibration, with a common interface for use ICalibration The NeuralCalibration reads the configuration file (monovariate/multivariate input)class uses a sigmoid trasnfer function on the hidden layer and an uses a linear function for output lyyer. The LinearCalibration use a linear coefficient to determine the pollutants estimation. the result is then processed in a non linear function for non linear sensibility correction, based on the temperature. NN estimator can take bot an array formed by temp and 2 co/no2/o3 electrodes for "monovariate" pollutant input, or temp and 6 electrodes for multivariate pollutants input. The library contains 2 junit test classes and a stub class for android.util.Log