20.500.12556/DKUM-16902 Blend.FL na Fakulteti za logistiko : diplomsko delo Blend.FL at Faculty of logistics Vseživljensko učenje je sodoben trend, ki se je v različnih oblikah poskušal uveljaviti že v zadnjih dvesto letih. Učni proces na FL UM poteka kot kombinacija avditornih vaj in predavanj ter študija s pomočjo tehnologije preko virtualne fakultete. Le-ta nudi študentu vse potrebne informacije za uspešen študij. Drugi del t.i. virtualnega učnega procesa so e-gradiva, ki so objavljena na spletu, dostop do njih pa imajo študentje FL z vpisom v portal s svojim uporabniškim imenom in geslom. V praktičnem delu naloge smo potrdili, da je sprotno delo študentov z e-gradivi nujen pogoj za njihovo uspešnost pri opravljanju predpisanih izpitnih obveznosti. Pripravili smo model, ki nam omogoča ločiti uspešne študente od neuspešnih na osnovi njihovega samostojnega dela v spletni učilnici. Model najprej napove višino končne ocene pri predmetu KML na osnovi povprečne ocene opravljenih kvizov. Na osnovi napovedi pa le-te klasificira v pozitivni in negativni razred. Model je zgrajen s konzervativnim pristopom, saj nobenega primera ne napove preoptimistično, to je v nobenem primeru negativne ocene, le-te ne napove kot pozitivno vrednost. Točnost klasifikacije je zelo visoka, saj napačno razvrsti manj kot 10% primerov. Lifelong learning is a modern trend, which was held in various forms in last two hundred years. Learning process at FL UM is a combination of traditional learning process, combined with e-learning system. In this way, student can get all necessary data. The other part of FL UM learning process is a web page with e-materials, which are placed online for the students and they can reach them with their username and a password. In the practical part we have confirmed that the ongoing work of students with e-materials is a necessary condition for their success in carrying out the prescribed exams. We have created a model that allows us to separate the successful from the unsuccessful students based on their individual work in the online classroom. The model first predicts the amount of the final evaluation at the course KML on the average estimate of quizzes. Based on the projections they classify the positive and negative class. The model is built with the conservative approach, because does not predict any case too optimistic, this means in any case of a negative assessment, they do not predict a positive value. Classification accuracy is very high, because it classifies wrong only 10% of the cases. ežgradiva napovedni model linearna regresija hibridno učenje prediction model linear regression hybrid learning true true false [S. Trafela] Slovenski jezik Angleški jezik Diplomsko delo 2010-11-30 16:10:53 2010-11-30 16:10:53 2022-04-13 09:00:47 2023-12-21 03:22:46 2010 0 Celje 0 0000-00-00 NiDoloceno NiDoloceno NiDoloceno 0000-00-00 0000-00-00 0000-00-00 378:004(043.2) 512256061 URN:SI:UM:DK:PPZBFRK8 UNI_Trafela_Sabrina_i2010.pdf UNI_Trafela_Sabrina_i2010.pdf 1 FB764513EB1AFAA14B0ED7137EE60A4D 18c850b3ef2ea318a7446f6c81893e111235e34096db8382668e38a0da59d335 c2148662-7c04-11eb-bb7a-00155d0001ca 20.500.12556/dkum/4ad7053b-d073-4da3-b29f-565384c1cdfb https://dk.um.si/Dokument.php?lang=slv&id=19384 Fakulteta za logistiko 0 0 0