WG creates solution to help reduce student dropout in the educational system

School dropout is the non-completion of a student period or an undergraduate course, in the case of higher education. Currently, it is one of the problems that most afflict educational institutions in general, being a phenomenon of multiple factors, which can occur with people from all socioeconomic and cultural contexts and teaching modalities.  

To help identify the possibility of dropout before it happens, the WG-LANSE is developing a technological solution for performing academic risk prediction services (dropout and/or failure) supported by a cloud infrastructure that pre-processes data, trains and executes machine learning algorithms. Selected in the RNP Advanced Services R&DI Program in 2021, the Working Group brings together researchers from the Federal University of Pelotas (UFPel), Federal University of Santa Catarina (UFSC), Federal Institute of Education, Science and Technology Sul-riograndense (IFSUL) and the Elimu Social startup.  

The tool is currently under development at the TRL4 stage, that is, at this moment the proof of concept is being put into practice, with application in an environment similar to the real one. The solution already has a cloud architecture that receives data from the Moodle environment and offers initial views of student interactions within the course, in addition to indicating possible student failure based on these interactions. 

“The solution is expected to help reduce dropout and failure rates for institutions that adopt it. The tests carried out with the prediction models point to the ability to anticipate dropout situations as early as the second week of a course with rates above 85% of accuracy”, informs Cristian Cechinel, WG coordinator and researcher at UFSC.  

Cechinel highlights RNP's participation in the project, which is allowing the transformation of research results on the prediction of academic risk into a product that can be used by teaching institutions that use Moodle in their teaching-learning processes. “We are working so that the MVP under development can be effectively offered to these institutions so that they have in their hands a tool that allows them to mitigate dropout and failure rates in their courses”, he says.   

According to the coordinator, the current expectation is to implement the solution that already exists in partner institutions (UFSC, UFPel and IFSUL), allowing its improvement through the collection of the experiences of the users of these institutions. “We also hope that the solution can be incorporated as one of the services to be offered in RNP's NasNuvens, as an ad-on that can be used by clients of Moodle hosting services”.  

The next stages of the project, which entered the second phase, involve moving forward with the development and implementation of the technology in order to reach the TRL9 maturity level (technology proven in an operational environment and with established production), in addition to the challenges of sustainability of the solution and of the business.