Problems of Indexing Classes of News Based on the Computed Importance of Words

Abstract

News is classified. Such classes as sports, politics, news on crime, gossips, business, etc., are common amongst newspapers in Nigeria. Interestingly most readers and patrons of newspapers adopt the rule of the thumb in choosing a suitable newspaper to read/buy. However, most newspapers try to cover all classes but end up in strikingly stress areas. This paper firstly explains the basic steps in generation of Document Indexes then secondly, highlights some of the problems associated with using full automation in classifications. Some of the problems identified are loss of relevance, loss of coverage and partial automation.

Keywords: Classification, Indexing, keywords