Machine Learning and Pattern Recognition

Abstract

Learning is needed when there is no human expertise existing or when human beings are unable to explain their expertise. In such a situation, one simply collects all the possible previous information, analyse it and then make a rule for future prediction or taking meaningful decision. When we plan to conclude such a work with the help of a computer by providing it ample amount of data and our past experience with tools and techniques, then the whole process becomes machine learning. Hence, machine learning can be defined as programming computers to optimise a performance criterion using example data and past experience. For example, recognition of spoken speech is being done by human beings seemingly without any difficulty, but cannot explain how they do it.

Defence Science Journal, 2010, 60(4), pp.345-347, DOI:http://dx.doi.org/10.14429/dsj.60.502