Practical efficiency evaluation method for bibliographical data classification

Abstract

Current state of computer science enables one to automatize routine tasks of increasing number in various spheres. One of such spheres is bibliographical data which still implies routine manual processing that causes human mistakes and considerable time waste. An important step for solving the problem is the development of bibliographical data classification algorithms as well as its efficiency evaluation methods. Existing evaluation methods consider the classification efficiency of each separate word while bibliographical data consists of homogeneous sequences that must be classified and evaluated together. This paper describes a new efficiency evaluation method that is suitable for bibliographical data classification efficiency evaluation due to it is taking into account not only the classification efficiency of each separate element but also the sequences of same class elements. Application of the developed method enables one to find the shortcomings of a classification algorithm and make conclusions concerning its practical applications.