Algoritmos de Clusterização e Python Científico apoiando Modelagem de Usuário

Abstract

Python is a interactive, interpreted, object oriented, multi-plataform programming language with libraries that aid the studies in Human-Computer Interaction throughout the use of machine learning algorithms, that assists in the construction of the Personas user type. These might help the developer during the interface development and may also enhance the ones that are currently in use. The moment that this technique may apply depends on how the information is acquired – with surveys or system's log. This paper presents an way to capture this information and transform this data, making possible to use clustering algorithms to group the collected profiles, enabling the automatization of the process in order to use in adaptive interfaces. Lastly, techniques grounded in statistics methods of spreadness measurements used to define a Personas and Python tools to graphically present the data, allowing an analytical comparison, will be demonstrated.