Preprocessing in noisy data of array sensor to enhance pattern recognition and classification problem Wahyu Hidayat, Agus Suroso
Theoretical High Energy Physics Division, FMIPA
ITB
Abstract
We will present the pre-processing of experimental data generated by a series of gas sensors, a device known as an electronic nose. We propose the use of a combination of statistical methods and artificial neural networks (ANN) to recognize data containing noise. As a result, we will get the optimal number of sensor arrays that can be used and also function properly. In addition, data pre-processing prove the increasing quality of data for classification and pattern recognition purposes.
Keywords: Noisy Data, Governing Equations, Array Sensor, ANN
