Machine Learning Approach for Error Handling - Linear Block Codes - A Case Study
摘要
The transferred information through an electrical communication channel is susceptible to additive noise with relatively, more certainty. Forward Error Correction mechanism facilitates information protection against channel noise through the overheads appended with the information prior to transmission. Redundancy introduced enables the receiver to retrieve the source information, against the effects of channel noise, at the cost of coding efficiency or code rate. Linear Block coding is such a Channel coding scheme, which possesses the feature that the linearly combined code words of the code result in another code word of the same code. Information word of length ‘k’ binary bits is transformed into a binary code word of length ‘n (=k+ r)’ bits with a one-to-one mapping, with ‘r’ number of overheads appended to ‘k’ bit information word in an (n, k) linear block code. Such a coding scheme with Hamming Distance of dmin can handle \( \frac{d_{min}-1}{2} \) number of effected bits in the received ‘n’ bit word. Information retrieval from the received (noise effected) is the functionality of the decoder/receiver, which can be very much implemented using Classification algorithms of Machine.