Classification of Error-Correcting Coded Data Using Multidimensional Feature Vectors

摘要

Error-correcting codes are used to encode data prior to transmission through noisy channel for their error-free data communication. In cryptographic secure communication, error-free key transmission is very essential for getting same key as used at transmitting end for message decryption at receiving end. As BCH codes, Golay codes, and Hamming codes are normally used in modern communication to encode data for error-free transmission over noisy channel, classification of distorted encoded data is an important activity and is very much required to decode intercepted data. We consider a statistical pattern recognition approach for classification of coded messages by applying multidimensional feature vectors and minimum distance criteria. The classification results obtained as shown in simulation results are quite encouraging, and we could classify such coded data with minimum 85 % and maximum up to 100 % success rate.