E-orientation system socio-psychological data sensitive

Abstract

Nowadays, the diversity of specialty choices in the different fields of study is a challenge for students. orientation has taken place at different levels of education and has become a crucial step in deciding the future career of students. In the majority of cases, the choice of the field of study is made subjectively, which sometimes results in failure or dropping out of educational institutions. In order to overcome this problem, we have proposed a specialty recommendation system that proposes the most appropriate field of study to students according to their academic and psycho-social background. In order to realize this recommendation system, we followed the construction methodology of recommendation systems, which is divided into three phases, starting with the first phase of data collection and preprocessing, during which we applied different techniques of data analysis, namely, principal component analysis for the reduction of the dimensionality of the database, followed by the second phase, the learning phase, whose objective is the construction of the students’ profiles through a clustering algorithm. The obtained profiles are used in the last phase, which is the prediction phase, where we have used a neural network to predict the appropriate recommendation. To validate our proposal approach, we developed a prototype using the Portuguese database that allowed us to analyze the relationship between different social data and the performance of students. The results obtained from the prototype are very interesting and reveal a strong correlation between the different types of data (numerical, psycho-social, or a combination of both) and the performance of students in a specialty.