Seminar on Computer Science and Applied Mathematics

PROGRAM

UTORAK, 04.04.2017. u 14:15, Sala 301f, MI SANU, Kneza Mihaila 36
Danica Vukadinović Greetham, The Department of Mathematics and Statistics, University of Reading
ANALIZA TVITERA: INDEKSI CENTRALNOSTI I IDENTIFIKACIJA GRUPA U TEMPORALNIM GRAFOVIMA
Postoji više načina da kreiramo dinamičku mrežu (ili temporalni graf) od mikro-tekstova objavljenih na društvenoj mreži Tviter. Koristeći adresiranje ('@') od oko 100 milliona prikupljenih poruka kreirali smo graf sa namerom da analiziramo kako se dinamika raspoloženja menja u zavisnosti od dinamike promena u strukturi grafa. Koristeći Katz-ov indeks centralnosti proširen na temporalne grafove pokazali smo da korisnici sa najvećim indeksom centralnosti šalju poruke sa pozitivnim sadržajem češće i sa negativnim sadržajem manje. Takodje, mere pozitivnog i negatvnog sadržaja poruka su nam omogućile identifikaciju stabilnih grupa aktivnih više meseci, gde veće promene u prosečnom raspoloženju grupe mogu da se vežu za eksterne dogadjaje. Na osnovu ovih opažanja, kreirali smo jednostavan model koji može da reprodukuje dinamiku slanja poruka i raspoloženja.
Bazirano na radu https://rsos.royalsocietypublishing.org/content/3/6/160162 [rsos.royalsocietypublishing.org] sa ko-autorima N. Charlton-om and C. Singleton-om sa Univerziteta u Redingu.


UTORAK, 11.04.2017. u 14:15, Sala 301f, MI SANU, Kneza Mihaila 36
Zoran Maksimović, Univerzitet odbrane, Vojna akademija
TWO METAHEURISTICS FOR SOLVING THE CONNECTED MULTIDIMENSIONAL MAXIMUM BISECTION PROBLEM
In this paper, a connected multidimensional maximum bisection problem is considered. This problem is a generalization of a standard NP-hard maximum bisection problem, where each graph edge has a vector of weights and induced subgraphs must be connected. We propose two metaheuristic approaches, a genetic algorithm (GA) and an electromagnetism-like metaheuristic (EM). The GA uses modified integer encoding of individuals, which enhances the search process and enables usage of standard genetic operators. The EM, besides standard attraction repulsion mechanism, is extended with a scaling procedure, which additionally moves EM points closer to local optima. A specially constructed penalty function, used for both approaches, is performed as a practical technique for temporarily including infeasible solutions into the search process. Both GA and EM use the same local search procedure based on 1-swap improvements. Computational results were obtained on instances from literature with up to 500 vertices and 60000 edges. EM reaches all known optimal solutions on small-size instances, while GA reaches all known optimal solutions except for one case. Both proposed methods give results on medium-size and large-scale instances, which are out of reach for exact methods.
This is joint work with J. Kratica, A. Savić


UTORAK, 25.04.2017. u 14:15, Sala 301f, MI SANU, Kneza Mihaila 36
Jasmina Perišić, Univerzitet Union Nikola Tesla
AN ONTOLOGY BASED FRAMEWORK FOR PERSONALIZED E-LEARNING IN MOODLE
In this work, we present a new approach to improve basic Moodle LMS (learning management system) by expanding it with two main functionalities: detection of learning style of the student and personalization of learning materials according to the student model. This approach is implemented through enriching Moodle with semantic technologies and ontology specially created to facilitate personalization. The result of applying these technologies is e-learning platform named MAL. In this paper, the functionality of the system and the way semantic technologies have been implemented into the MAL platform will be presented. We also conducted evaluation study to compare the effectiveness of both the non-adaptive approach and the adaptive approach based on semantics. The evaluation has been carried out at the Faculty of Entrepreneurial Business, the University Union "Nikola Tesla", Belgrade. The results are promising since they indicate that semantic adaptation of learning content implemented in MAL leads to significantly better learning outcome and improved efficiency of the learning process compared to the non-adaptive e-learning courses. They also indicate a positive attitude of students towards the developed learning environment.