Full text
| |
| Author(s): |
Joaquim Jose Fantin Pereira
Total Authors: 1
|
| Document type: | Master's Dissertation |
| Press: | Campinas, SP. |
| Institution: | Universidade Estadual de Campinas (UNICAMP). Faculdade de Engenharia Elétrica e de Computação |
| Defense date: | 2007-06-07 |
| Examining board members: |
Takaaki Ohishi;
Rosangela Ballini;
Paulo Sergio Franco Barbosa;
Ricardo Ribeiro Gudwin
|
| Advisor: | Takaaki Ohishi |
| Abstract | |
Decision making, in any area and in many different levels, is a process with growing complexity, mainly if you consider the level of uncertainty related to the future. In this context, the possibility of forecasting plays a major role in an efficient decision. On the other hand, pattern recognition tools are important in many areas, like fitting typical behaviors and in control systems, as well. In this context, we propose a visual programming language, called VisualPREV Language, intended to make easier the conception and execution of forecasting and pattern recognition models. Within this language, visual blocks that can be put into a diagram (computational visual interface) represent concepts involved when modeling the processes. These models can be configured, executed and stored for future access. Although these approach implies losing exclusive advantages of traditional programming (like flexibility of generic programming, for example), VisualPREV decreases considerably the amount of time needed for creating specific models for forecasting and pattern recognition. In few applications with relevant data, the language was evaluated based on usability metrics, and the results were discussed throughout the text (AU) | |