A FORMAL MECHANISM FOR AUTOMATIC CLASSIFICATION OF LEARNING OBJECTS

This article is mainly developed a formal mechanism for automatic of a set of learning objects according to the value of some indicators in order to maintain backward in a continuous cycle of improvement. This involves building a measure space, a measure function, a function of dissmilarity and a method of classification. in addition to this, we applied a measurement process that helps to give greater meaning to the information obtained for each subject and the possible clusters obtained as a result of the classification. This mechanism allows classifying learning objects of different types, tested with different sets of indicators, including sorting achieved independently of the meanings of the latter. The classification method fails to do so in linear time, which can handle a large number of objects without significantly increasing classification time. It does define an efficient procedure for adding or removing objects form the classification, with the goal of eliminating errors in the evaluations of these objects or to update these values because the object has been Improved. Se proposes that objects are evaluated by a team multidiscipilnary, in order to evaluate different aspects. It is also argued as the mechanism meets a number of key characteristics for rating systems. Finally it shows how the adddition or removal of objects can make objects change learning a cluster to another, or even disappear clusters.