Linear block code are an important class of error correction code which are used to correct errors especially in digital communication and data storage systems. Error correction code aim to reduce the probability of bit errors. The decoder of linear block code usually uses the syndrome calculation method and the coset leader placement to determine the error pattern. Supervision is carried out by matching the syndrome with the error pattern by means of a look-up table. This method is called syndrome decoding.

This paper will present an alternative method for decoding linear block code by using a back propagation neural network. The back propagation neural network is trained in order to recognize the input pattern in the form of the linear block code (7.4) and the target in the form of the original message.

The network that has been trained is tested with 1000 bits of random data. The channel simulation used is Binary Symmetric Channel (BSC) with various error probability values. Based on observations, linear block code with artificial neural networks decoder can reduce the Bit Error Rate (BER) by an average of 29.5%