Réinitialiser mot de passe

Experimental Evaluation of a Dynamic Cubing System: Workflow, Metrics, and Prototype
In EDA 2015, vol. RNTI-B-11, pp.177-192
This article originates in the efforts we made in the prototyping stage
of a project conducted in the field of new agile BI applications. In this earlier
work, we defined a model and algorithms for handling dynamic cubes that are
updated as single fact comes in. The proposal includes the definition of a tree
structure, called DyTree, used to store these cubes. A comprehensive experimental
evaluation was conducted to observe the behavior of the proposal under
some varying circumstances or settings, to help in understanding its working
and to provide feedback to improve the solution. In this paper, we describe the
workflow which has been used for this experimental evaluation. Performance
metrics and behavioral metrics are defined as output of the experimental evaluation
workflow. Different types of data sets and parameters for algorithms are
the input used to configure the experiments. Using this environment and the prototype,
some examples of experimental studies are presented. This shows how
different scenarios can be simply constructed and how the experimental results
can be used to understand the behavior of our proposal and to evaluate performance.