Ponente en el Congreso Nacional de Ciencia y Tecnología

Fecha: 
17/01/2017
Fecha de finalización: 
20/01/2017

Hierarchical Cluster Analysis is used in clustering, prediction, and decision-making processes. However, in cluster analysis techniques it is difficult to test assumptions, implement the method and interpret the results. In order to enrich the hierarchical clusters techniques, this document proposes a new technique that takes advantage of the noise resistance, speed of calculation and results of easy interpretation that are characteristic of the statistical implicative analysis. To demonstrate how the implicit statistical analysis can be used in hierarchical cluster analysis, the authors have worked with images whose groups are simple to identify and value. This research carried out experiments, randomly grouping images between 2 and 63. These images were grouped hierarchically using two software and then the 35 students read and interpreted the results. Each student indicated the level of agreement on how the images were grouped, then a hypothesis test was performed on the sample of possible clusters to infer the global results. Approximately 70% of the students agree or strongly agree with the groups created with the implicate statistical analysis