Naslov
Determining individual learning strategies for students in higher education using neural networks

Autori
Kliček, Božidar ; Oreški, Dijana ; Divjak, Blaženka

Izvornik
International journal of arts & sciences (1557-718X) 3 (2010), 18; 22-40

Vrsta, podvrsta i kategorija rada
Radovi u časopisima, članak, znanstveni

Ključne riječi
Neural network; academic success; learning strategy; personalized learning; discriminant analysis; multiple regression

Sažetak
This paper aims to predict the student success identifying factors that lead to success and use this ability to determine the best individual learning strategy for the university students. This model covers the whole process of education, including choosing the proper faculty, recognizing most fruitful behavior and learning style for particular student and choosing the most useful courses. Using data taken from graduated students at the end of their study, this research investigates data mining technique, neural networks, as tool to identify those predictors. Analysis indicates that neural network models are capable of modeling nonlinear fuzzy mappings presented in the data and may offer some advantages over traditional statistical models in this domain. Neural network model revealed there is a significant degree of diversity within the students and learning process is quite personalized. Thus standardized set of guidelines could not be constructed and general strategy could not be applied for all students, yet several groups have their own ways to determine learning strategy. Paper emphasis the importance to recognize that a “one size fits all” model can not effectively serve the needs of all students neither their institutions. So rather than trying to stick every student into one model, the answer is to provide different models for different groups of students and offer flexibility to fit the needs of the students. Neural networks have potential to serve as an effective tool in determining learning strategy, which can be used as tools for students advisers and an on-line application for the students. Generic architecture of facility collecting data, discovering regularities in data with neural networks and interface for advising services for the advisers and students is finally described.

Izvorni jezik
Engleski

Znanstvena područja
Informacijske i komunikacijske znanosti

POVEZANOST RADA