Ion-channel sensors can be used for detection of biochemical reagents. A silicon-based ion-channel platform has been developed for stochastic sensing of molecules. In this project, we develop techniques to classify signals from Ion channels to detect the presence of binding molecules or ions. We first explored the traditional methods like parametric modeling using Autocorrealtions and Spectral Density. In such a model we assume the Ion channel can be described by n-Markov states and parameters alpha and beta determine the switching between states. We used some of these concepts and developed an exercise for EEE598 class Bioelectronics - cells and Systems. We then explored the machine learning approach which showed promising results. We present here techniques to extract appropriate features from sensor data using a combination Walsh-Hadamard Transform and Principal Component Analysis. We use neural network techniques to discriminate between the analytes.