Model@ by UPC

The objective of Model@ is to implement an automatic system that can collect data from the UOC’s datamart and generate predictive models from it.

These data extracted from the UOC datamart are processed with ETL tools, using different grouping criteria, in order to generate the necessary data models to be used in the calculation of the predictive models.

A critical feature of the project is the elasticity of the system: through different definitions in files, it is indicated which are the data to be collected from the datamart and what is the structure of the models. This ensures that all configurations can be dynamically modified without having to change the code.

For example, one of the implemented models analyzes information related to the student’s activity to make predictive estimates around the possibility of success in their assessment.

The inLab is responsible for processing the data and preparing it for input into the predictive model and running the models specified by the UOC.