Створення рекомендаційної системи підтримки прийняття рішень для запису на вибіркові навчальні дисципліни
Loading...
Date
2017
Authors
Горборуков, Вячеслав
Олецький, Олексій
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
У статті розвинуто підходи до побудови рекомендаційної системи підтримки прийняття рішень на базі автоматизованої системи запису студентів на вибіркові навчальні курси, що була
розроблена в Національному університеті «Києво-Могилянська академія». Сформульовано відповідні оптимізаційні задачі, при цьому основний акцент робиться на пошуку методів знаходження
адекватних цільових функцій. Розглянуто можливість застосування евристичних методів, зокрема
на базі PageRank-подібних методик та методу Сааті. Крім того, розглянуто задачу автоматизованого розрахунку кількості груп в умовах нечітко сформульованих м’яких обмежень.
The paper regards approaches to developing a recommendation decision making supporting system for enrollment in selective educational disciplines. Reciprocal optimization problems including multi-level and multi-criteria problems of choosing sets of disciplines are formulated. Hereby it is taken into account that optimization criteria and goal functions might be defined not well and therefore should be estimated, e.g. by taking into account the available experience. These functions may be non-additive and non-linear, so combining functions of getting usefulness of sets by usefulness of single disciplines should be involved. Some approaches to getting such combining functions are proposed in the paper. Possible ways of taking into account possible connections between disciplines (e.g. if one discipline should rely upon another one and therefore should precede it) which form the framework of a curricula template are also regarded. A possible fragment of such a graph of relations in a Prolog-like notation is given. The ontology of topics covered by disciplines is considered. A possibility of applying heuristic approaches, such as PageRank-like techniques and the Saati method based on pair comparisons, has been explored. As a result, we regard a possible generative model for getting transitional probabilities and some equations for getting unknown coefficients. The problem of automated calculating of amount of groups and forming groups of students under soft and fuzzy constraints is regarded as well.
The paper regards approaches to developing a recommendation decision making supporting system for enrollment in selective educational disciplines. Reciprocal optimization problems including multi-level and multi-criteria problems of choosing sets of disciplines are formulated. Hereby it is taken into account that optimization criteria and goal functions might be defined not well and therefore should be estimated, e.g. by taking into account the available experience. These functions may be non-additive and non-linear, so combining functions of getting usefulness of sets by usefulness of single disciplines should be involved. Some approaches to getting such combining functions are proposed in the paper. Possible ways of taking into account possible connections between disciplines (e.g. if one discipline should rely upon another one and therefore should precede it) which form the framework of a curricula template are also regarded. A possible fragment of such a graph of relations in a Prolog-like notation is given. The ontology of topics covered by disciplines is considered. A possibility of applying heuristic approaches, such as PageRank-like techniques and the Saati method based on pair comparisons, has been explored. As a result, we regard a possible generative model for getting transitional probabilities and some equations for getting unknown coefficients. The problem of automated calculating of amount of groups and forming groups of students under soft and fuzzy constraints is regarded as well.
Description
Keywords
автоматизація, e-learning, рекомендаційна система, оптимізаційна задача, оптимальний вибір множин, automation, recommendation system, стаття
Citation
Горборуков В. В. Створення рекомендаційної системи підтримки прийняття рішень для запису на вибіркові навчальні дисципліни / Горборуков В. В., Олецький О. В. // Наукові записки НаУКМА. Комп'ютерні науки. - 2017. - Т. 198. - С. 54-58.