The errors that may occur in transmitted numerical data on a channel affected by burst errors, are corrected via the operations of: ordering the numerical data in blocks each comprising a definite number of data packets; generating for each block a respective set of error-correction packets comprising a respective number of correction packets, the respective number identifying a level of redundancy for correcting the errors; and modifying dynamically the level of redundancy according to the characteristics of the bursts and of the correct-reception intervals between two bursts. Preferential application is on local networks, such as W-LANs for use in the domestic environments.