S-SAM: a semantic self-adapted method for categorizing annotated resources

Abstract

The present final degree project proposes a new method for automatic classification of resources labelled with tags coming from a folksonomy of social tagging systems. It is the result of a variation of SAM, a self-adapted method, that is, a method of automatic classification of annotated resources, which have been done by some researchers of the Public University of Navarra. The method, called S-SAM (or Semantic SAM) have as their goal to improve the classification of annotated resources by means of this automatic method, without using human force, in order to make more accurate the knowledge representation and information recovery. To do so, it has been chosen the final degre project of Ciordia, 2011 as a pattern to follow in the implementation of SAM and S-SAM, which is a Java program that needs some data allocated in MySQL format databases. The research is divided into two parts. The first part studies the way a subset of resources is classified using the number of occurrences versus using the fitness of the annotation (that is, a consensus evaluation from experts). The second part also studies this but using the whole set of resources (all the annotations). Once obtained the results, they will be compaired finding out which way classifies the best.