Probabilistic Reasoning still the Shortest Path in Many Problems

While the notion of chance may have been inherent in us from the beginning of the thinking age, it is a gambler's dispute in the 17th century that led to the creation by two famous French mathematicians of one of the most powerful mathematical theories, the theory of probability. It has since helped solve countless problems. In engineering, it is not the existence of uncertainty that is disputed, but how to deal with it. Probabilistic thinking becomes cumbersome when too many sources of variability exist. Engineers and physicists resort to "Black Box" models using data algorithmic approaches. Such is the case in some classification problems brought out by pattern recognition. We look at a character recognition problem, and present a probabilistic solution that is simple and easy to implement.