Hybrid Event

Event program

Thursday, 5/25/2023 3:00 PM - 6:00 PM, Congress hall, Grand hotel Adriatic, Opatija 3:00 PM - 6:00 PM Papers  1. M. Fertalj, L. Brkić, I. Mekterović (Faculty of Electrical Engineering and Computing, Zagreb, Croatia) Metrics for Estimating Accuracy, Reliability and Bias in Peer Assessment  The challenge of teaching staff being overworked and not having enough time to provide individualized feedback and guidance to students is a significant problem. Open-ended assignments foster creative thinking but add to the workload. The abundance of solutions makes individualized assessment necessary. Peer assessment, where students evaluate their peers' assignments based on their motivation, knowledge, and resourcefulness, has proven effective in higher education courses. This paper reviews metrics used in peer assessment, defines what constitutes correct and consistent evaluation, and suggests ways to choose appropriate metrics for learning objectives. The paper also highlights limitations of current metrics, seeks to enhance accuracy in peer assessment, and proposes derived metrics based on existing ones to increase consistency and fairness.  2. A. Naumoski, G. Mirceva, K. Mitreski (Faculty of Computer Science and Engineering, Skopje, Macedonia) Exploring the Relationship between Indoor Playrooms and Population in Skopje  Play has a crucial role in childrens’ growth. It teaches them new skills, strengthens their self-confidence, and promotes creativity. Both indoors and outdoors offer play opportunities, but a nearby playground is crucial as it provides a safe and fun place for children to exercise and explore. This can lead to a healthier lifestyle and improved social skills through interacting with other children. In this paper, we analyze the spatial relationship between indoor playrooms and the population in the study area in Skopje. Our study is based on publicly available data on indoor playrooms, including their location, user satisfaction ratings from customers, and data on population density in the area. Our goal is to find potential locations for new indoor playrooms and improve existing indoor playroom offerings through interpolation, hot spot analysis, and spatial data analysis with multiple ring buffers. Our analysis reveals spatial areas of high population density that offer opportunities to improve indoor playroom products and services. 3. M. Begović, E. Avdagić-Golub, B. Memić, A. Kosovac (Faculty of Traffic and Communications, Sarajevo, Bosnia and Herzegovina) Predicting Customer Behavior in Support Channels Using Machine Learning  Support channels represent unique opportunity to improve customer satisfaction by offering a consistent experience in resolving customer issues. Several surveys show that customers have raised their standards of customer support services. While only a few years ago customers willingly waited a long time to speak with one of the service agents and were patient for their problem to be resolved, today's customers have very limited patience and want a solution to the problem immediately. Customers don't want to settle for a mediocre support channel experience. Support channels must provide superior service capacities, so that customers see that the company values their choice and time. Efficient management of support centers implies accurate modeling of customer behavior on hold. The subject of our research is the application of data research techniques to the prediction of customer behavior in support channels. In this paper, we apply machine learning methods to predict customer behavior. Based on historical data in the service system, we use classification algorithms to predict customer patience in service channels. 4. J. Krčadinac, E. Maruševec, L. Jerković, I. Kovač, J. Zloić, A. Šarčević, M. Vranić (Faculty of Electrical Engineering and Computing, Zagreb, Croatia) Modeling Tennis Matches Using Monte Carlo Simulations Incorporating Dynamic Parameters  Although it may seem to be one of the more unpredictable sports, tennis can be modelled quite accurately using the Monte Carlo method because of its rules. The purpose of this study is to test how successful is a Monte Carlo simulation that incorporates dynamic parameters of tennis at predicting the winner of a given match. A regular Monte Carlo simulation that predicts the winner of a tennis match only needs two parameters, the percentage of all service points won for each player, while our more advanced method also considers how fatigue effects a player’s performance and the percentage of winning service points after “breaking” his opponent’s service game. Our study consisted of calculating the relevant statistics from a large dataset of tennis matches, making both the regular Monte Carlo simulation and a simulation that incorporates dynamic parameters and implementing them using previously calculated player statistics and a different dataset of matches for predicting the winner. After successfully simulating more than 1200 matches, we determined that both methods are almost equally accurate when it comes to predicting the winner, with the percentage of correct guesses being around 60%. When the simulations are more than 70% sure that one player will be the winner, the percentage of correct guesses jumps to 67%. 5. M. Fabijanić, D. Ružak (Algebra University College, Zagreb, Croatia), A. Novosel, T. Hlupić (Poslovna inteligencija d.o.o., Zagreb, Croatia) Data Warehouse-Based Analytical System in Private Higher Education Institution  Data warehouses are present in the business environment for several decades, providing businesses with deeper insight into their own business data through which they gain competitive advantage and decision-making support. However, the integrated characteristics of data warehouses can be a key factor in improving process automatization through the integration of various heterogeneous sources, speeding up certain processes, and providing process integration in the organization without impacting the existing infrastructure. In this paper, an analytical system based on a data warehouse is presented which integrates data sources in a high education institution, such as internal operational applications and structured and semistructured files, and incorporates new data sources such as websites and their interfaces. The process which collects and quantify data from web pages about published conference papers of education institution personnel will be shown as an example. Through introducing the integration of the data sources, processes that share the same data scattered across multiple files or need multiple manual entries gain the possibility of using a single source of truth, while also having the possibility of automatic integration of open-access data. In the paper, the necessary processes are identified and their respective datamarts are proposed, providing proof-of-concept for several chosen processes. 6. B. Wang (Chengdu City Construction Investment & Management Group Co., Ltd, Chengdu, China), Y. Wang, Z. Wang (Wuhan University, Wuhan, China) Opinion Mining on E-commerce Live Broadcast of Agricultural Products Based on WLDA-Apriori Model  With the development of the popularization of information technology in rural society, the e-commerce live broadcast of agricultural products has provided an impetus for agricultural marketing from rural to urban areas. But as an emerging industry, it has problems in live broadcast content, marketing strategies, market supervision, industry norms and agricultural aid effects. In order to fully understand public opinion and put forward suggestions on the sustainable development of agricultural products e-commerce live broadcast, this paper proposes an opinion mining method based on WLDA-Apriori model. The method combines topic model with association rule algorithm, which can further analyze the implicit semantic relations among topic words. This paper also conducts the empirical study with related social media data from Weibo, and obtains eight topics the Chinese public concerned. Besides, the public opinion are identified and discussed from aspects of Industry Development, Market Effect, Live Broadcast Subject and Negative Effect, and corresponding suggestions are put forward. The results of this paper are more universal, and the suggestions are consistent with the needs of the public. But this paper does not integrate and analyze social media data from multiple platforms, nor does it analyze public opinion from the sentiment level. 7. V. Perković, F. Belac, K. Boroš, M. Kos, L. Humski, D. Pintar (Fakultet elektrotehnike i računarstva, Zagreb, Croatia) CroStats - Visualization of Population in Croatia  From time to time, population census in Croatia becomes an actual topic in conversations. When it does, it usually shows abstract numbers and an occasional graph on the evening news. After doing some research, we couldn't find a way for the average person to get a closer look at the results. At least not without comparing infinite amount of data from the table. Solution to that problem is CroStats – a simple web interface for the population census data visualization. Main feature of the CroStats is an intuitive and simple graphical display. It analyzes most important categories of the population census such as population, age, birth rate and mortality by county. It is shown on an interactive map of Republic of Croatia. Extra features are graphs of changes by year and interesting facts about counties that the broader audience may not know. 8. V. Fomichov (Moscow Aviation Institute, Moscow, Russian Federation) A New Approach to Semantic Parsing of Metonymic Phrases by a Business Intelligence System  The functioning of numerous business intelligence systems (BIS) substantially depends of the content of databases. In many thematic domains, this content is highly dynamic, e.g., in case of the prices for oil, aluminum, wheat, etc. It is the reason for the birth of a BIS project aimed at updating the content of relational databases by means of extracting actual values of data from the texts in natural language (NL). The principal subject of the paper is theoretical foundations of building semantic representations (SRs) of the metonymic phrases. In linguistics, a metonymic construction is a linguistic construction in which an entity is referred to by the name of something closely associated with that thing or situation. For instance, a piece of sport news may not call explicitly the hockey team achieved the victory in the contest but indicate the color of this team’s emblem or indicate a thing depicted on this team’s emblem. The main scientific results set forth in the paper are as follows: (a) a formal model of a knowledge base helping to fulfil semantic parsing of metonymical phrases; this model expands the theory of K-representations (knowledge representations); (b) an original algorithm of building SRs of metonymic phrases; (c) the structure of a BIS implemented this algorithm and updating the content of a relational database. 9. S. Sharma, B. Stackpole, J. Pelletier (Rochester Institute of Technology, Rochester, United States) A Common Pentest Output Schema for Business Intelligence System Ingestion  Data-driven Business Intelligence process improvements demand proactive digital vulnerability discovery and exploitation enumeration. These risk discovery and analyses practices are commonly known as penetration tests (pentests), and have emerged as requirement for most organizations under many compliance regimes such as the global Payment Card Industry Data Security Standard (PCI-DSS) and the federal Gramm-Leach Bliley Act (GLBA) in the United States. Though a growing body of research addresses the utility of pentests as a valuable Business Intelligence method, pentest outputs are not standardized and require time-consuming generation and ingestion before they can provide inputs to Business Intelligence Systems. In this study, we examined several pentest reports and propose a standardized schema for output generation in a common format that is easy for Business Intelligence Systems to ingest. We propose that this improvement will allow more rapid delivery of accurate risk information to executives, managers, and information technology professionals. We also believe that this standardized schema using a common output datatype will further ease the translation of cybersecurity risk information into business impact analyses and organizational risk registers. 10. M. Fabijanić (Algebra University College, Zagreb, Croatia), I. Mekterović (Faculty of Electrical Engineering and Computing, Zagreb, Croatia) Partial SQL Query Assessment  Automated grading systems in education have been around for sixty years. They have found applications in areas such as online learning systems where virtually an unlimited number of users can test their knowledge that would not be possible to evaluate manually. Implementations within Massive Open Online Courses are a good practice in which users have the opportunity to self-test, and get instant feedback, making the learning process more efficient. Within universities, automated grading systems allow teachers to evaluate solutions and provide feedback for thousands of submissions in a short amount of time. The time frame to complete the assessment process is often short, and automated assessment systems give teachers more time to focus on the quality of exams and exercises. This paper presents an overview of methods used in automatic SQL query evaluation systems, from early implementations when the goal was only to evaluate solutions binary, to today when new technologies enable functionalities such as partial and configurable evaluation, rich and customized feedback, learning analytics, which helps to detect learning patterns, assess code quality, detect plagiarism, etc. These methods and technologies are not exclusive, and combining different approaches could make an automated grading system more comprehensive and applicable.

Basic information:

Boris Vrdoljak (Croatia), Mirta Baranović (Croatia), Matteo Golfarelli (Italy), Mihaela Vranić (Croatia)

Marko Banek (Croatia), Mirta Baranović (Croatia), Ladjel Bellatreche (France), Ljiljana Brkić (Croatia), Alfredo Cuzzocrea (Italy), Todd Eavis (Canada), Matteo Francia (Italy), Enrico Gallinucci (Italy), Dragan Gamberger (Croatia), Matteo Golfarelli (Italy), Marko Gulić (Croatia), Igor Mekterović (Croatia), Damir Pintar (Croatia), Sandro Skansi (Croatia), Zoran Skočir (Croatia), A Min Tjoa (Austria), Mladen Varga (Croatia), Mihaela Vranić (Croatia), Boris Vrdoljak (Croatia)

Registration / Fees:

The discount doesn't apply to PhD students.

Boris Vrdoljak
University of Zagreb
Faculty of Electrical Engineering and Computing
Unska 3
HR-10000 Zagreb, Croatia

Phone: +385 1 6129 532
Fax: +385 1 6129 915
E-mail: boris.vrdoljak@fer.hr

The best papers will get a special award.
Accepted papers will be published in the ISSN registered conference proceedings. Presented papers in English will be submitted for inclusion in the IEEE Xplore Digital Library.
.............
There is a possibility that the selected scientific papers with some further modification and refinement are being published in the following journals: Journal of Computing and Information Technology (CIT), MDPI Applied Science, MDPI Information Journal, Frontiers and EAI Endorsed Transaction on Scalable Information Systems.