Framework dirigido por datos y modelos para evaluar la calidad en uso del Software

Resumen

The interaction between users and computer systems has changed in recent years due to the proliferation of new devices and the emergence of new interaction paradigms. Examples of this are ubiquitous computing and context-aware systems, which create environments in which services react to user conditions to adapt to their needs. Also, these developments have led to the generation of large amounts of data containing information about user interaction, activity, and behavior. Such data can contain relevant information about the use of applications and the user's perception regarding the interaction with those. Thus, the analysis of these data with the appropriate techniques could allow the estimation of the user experience based on the patterns that are exhibited. In this work, we study the automatic evaluation of the quality of the interaction from activity data. The model of Quality in Use is considered to assess the effect of interaction with a software application and its adaptation to the user's context. In our research, we propose that it is feasible to design an environment that can be adapted and integrated into context-aware systems to ensure and optimize the Quality in Use. For the proposed research objectives, a framework is defined to evaluate the Quality in Use based on interaction data and models. In this framework, we propose a set of heuristic metrics, intended to estimate quality, and user context models, based on ontologies and probabilistic structures, intending to represent the elements of the interaction. Through Data Mining techniques, quantitative and qualitative results are produced to assess the Quality in Use. The proposed framework and its integration are validated through case studies and in the industrial environment of a software company.