Search

Population census is often a popular topic in everyday conversations. After extensive research, a conclusion was made that currently no services exist that offer a convenient, easily interpretable information about the census. Our solution for this problem is CroStats – a simple web interface for population census data visualization intended for general public. The central feature of the CroStats is its intuitive and simple graphical display. It analyzes the most important categories of the population census by county such as population, age, birth rate and mortality. All this information is displayed on an interactive map of the Republic of Croatia. Extra features offered are graphs of changes by year and interesting historical and demographic facts about counties that the broader audience may have interest in.

The Sava River Basin, as the largest river basin in Croatia, is of great importance for the water resources management and hydrological research in Croatia. The general intensification of extreme climatic conditions increases the number of flood events worldwide, which may indirectly affect the morphodynamical behavior in the riverbed. Valuable information on the flood-generation mechanisms may be contained in the flood hydrograph shapes, which can be expressed as continuous functions using a functional data approach [1]. In this research, we performed a preliminary analysis of annual flood events at the gauging station (GS) Zagreb by applying a clustering mechanism based on functional data to identify a set of representative hydrograph shapes that contain valuable information for future research on scour in the vicinity of bridges. Flood events that capture river’s flood regime can be represented as classical multivariate data or functional data, to be subsequently classified by applying a clustering algorithm. The functional analysis can provide a more objective and reproducible definition of the actual hydrological phenomena than the classical multidimensional analysis because it avoids the subjective selection of a set of hydrograph characteristics. Functional data analysis starts by considering the discrete data observations as part of a finite- dimensional space spanned by a set of basis functions and coefficients that define their linear combination. In addition, the choice of the appropriate order of polynomial segments and the number and placement of the knots are determined to represent the main characteristics of the hydrograph shapes. Finally, the resulting set of coefficients is used as an input for clustering. Since the hydrograph shape clusters can only be validated by their interpretability and usefulness, their evaluation was performed graphically [1, 2]. The input data for the analysis was a historical time series of daily discharge data from the Zagreb gauging station in the Sava River basin for the reference period 1960-2019, provided by the Croatian Meteorological and Hydrological Service. The time series consists of 21878 observations (60 years of data). The annual maxima method was applied to the time series to extract independent flood events corresponding to the maximum annual peak discharge values. For each event the flood volume and duration were determined over a fixed window corresponding to the longest duration of the rising and falling limb of all considered flood events, resulting in 16 days before and 38 days after the peak discharge (55 daily observations). The baseflow was separated from the direct flow using the R package “lfstat”, with the direct flow component normalized by dividing the ordinate of each hydrograph with the total volume of the flood event. Finally, a functional data analysis was performed by spanning B-spline functions of differing ranks onto the resulting direct flow data observations followed by implementing a k- means clustering algorithm to separate flood events into different clusters. For each cluster, a median was determined to represent three distinct types of hydrograph shapes. For this analysis the B-spline rank of 25 and the K value of 3 were ultimately chosen. This approach resulted in identifying three distinct types of flood events: (1) slow events described as both elongated rising and falling limbs, (2) intermediate events with steep rising limbs and moderately steep falling limb, and (3) fast events with steep rising and falling limbs of the hydrographs. This preliminary analysis was conducted in order to exploit the valuable process information contained in flood hydrograph shapes, which can be of great value for the future research within the R3PEAT project (“Remote Real- time Riprap Protection Erosion Assessment on large rivers”, UIP-2019-04-4046) supported by Croatian Science Foundation, that explores influences on the riverbed erosion around structure of bridges crossing large rivers in Croatia [3].

Sustavi za društveno umrežavanje (Facebook, Twitter, Instagram i slični) široko su rasprostranjena digitalna platforma za komunikaciju i interakciju milijardi ljudi. Podatci o interakciji ljudi pohranjuju se na strukturiran – računalno čitljiv način, što (uz odgovarajuću suglasnost ljudi o čijim se podatcima radi) omogućuje njihovu obradu, analizu i izlučivanje različitih zaključaka. Dostupnost podataka o interakciji temelj je svih istraživanja društvenih mreža – kako onih sociologijskih i psihologijskih, tako i onih iz područja računalne znanosti. U ovome doktorskome radu provedena su društvena istraživanja kroz koja su (uz suglasnost ispitanika) prikupljeni podatci o interakciji korisnika na sustavu za društveno umrežavanje Facebook te usporedno korisničke procjene njihova odnosa s ljudima s kojima na Facebooku interaktiraju. Time je oformljen referentni podatkovni skup za istraživanja društvenih mreža. Oformljen referentni podatkovni skup u radu je detaljno analiziran. Korištenjem referentnog podatkovnog skupa, osmišljen je, implementiran, ugođen i verificiran model za računanje društvene udaljenosti na temelju podataka o interakciji ljudi na Facebooku. Za razliku od prethodnih istraživanja koja su intenzitet odnosa među ljudima prikazivala primjenom koncepta težine ili snage veze, društvena udaljenost inovativan je način iskazivanja intenziteta odnosa među ljudima kroz primjenu koncepta udaljenosti. Osnovna značajka koncepta udaljenosti jest da se veći intenzitet prikazuje manjim brojem, što je vrlo pogodno za daljnje primjene kao što su određivanje vremena propagacije informacije ili određivanje približne geografske lokacije korisnika. U radu je definiran model obogaćenoga društvenoga grafa. Obogaćeni društveni graf formalni je prikaz društvene mreže, tj. odnosa među ljudima. Obogaćenim društvenim grafom moguće je zapisati različite vrste i intenzitete odnosa među ljudima. Intenzitet odnosa među ljudima bilježi se kao njihova društvena udaljenost, a ona se računa na temelju modela koji se u ovom radu uvode. Social networking sites (Facebook, Twitter, Instagram, etc.) are widely used digital platforms for communication and interaction between billions of people. Data about users’ interaction is stored in a structured and computer readable format, which makes possible (with the appropriate consent of the people whose data are) to process it, analyze and come up with various conclusions. The availability of data about people’s interaction is the foundation for all social networks' researches – both sociological and psychological, as well as those in the field of computer science. In this doctoral thesis a number of social surveys were conducted (with the appropriate consent of the examinees) and reference datasets for social network researches were collected. Datasets contain data about users’ interaction on the social networking site Facebook and users’ estimation of the relationship intensities with their Facebook friends. Collected reference datasets are analyzed in detail. By using a reference dataset, a model for computing social distance based on people's interaction data on Facebook is designed, implemented, fine-tuned and verified. Unlike previous researches showing the intensity of interpersonal relations by applying the concept of weight or tie strength, social distance is an innovative way of expressing the intensity of relationship between people through the concept of distance. The main feature of the concept of distance is that the higher intensity is shown by the smaller number. It is very suitable for further applications such as determining the propagation information time or determining the user’s approximate geographic location. Thesis defines model of enriched social graph. Enriched social graph is formal representation of social network, i.e. formal representation of relations between people. Enriched social graph can record different types of interpersonal relations and their intensities. Intensity of interpersonal relations is shown as their social distance, which is calculated by using model previously introduced in this thesis.

Predavanje se bavi izgradnjom i analizom društvenih grafova. Prezentirani su primjeri iz telekomunikacijske i financijske industrije te primjer izgradnje društvenog grafa na temelju interakcije korisnika na društvenoj mreži Facebook.

Facebook, the popular online social network, is used on daily basis by over 1 billion people each day. Its users use it for message exchange, sharing photos, publishing statuses etc. Recent research shows that a possibility exists of determining the connection level (or strength of their friendship, tie strength) between users based on analyzing their interaction on Facebook. The aim of this paper is to explore, as a proof of concept, the possibility of using a model for calculating strength of friendship to compare and classify ego-user’s Facebook friends. A survey, which involved more than 2500 people and collected a significant amount of data, was conducted through a developed web application. Analysis of collected data revealed that the model can determine with a high level of accuracy the stronger connections of ego-user and classify ego-user’s friends into several groups according to the estimated strength of their friendship. Conducted research is the base for creating an enriched social graph – graph which shows all kinds of relations between people and their intensity. Results of this research have plenty of potential uses, one of which is specifically improvement of the education process, especially in the segment of e- learning.

Tema ovog rada bilo je stvoriti model za predviđanje rezultata kuglačkih utakmica, a pritom naučiti više o analizi zadanog podatkovnog skupa te baratanju istim. Cilj je svakako bio ostvariti predviđanja koja su približna stvarnima i za pojedinačne rezultate i za ekipne. Korišteno je prediktivno modeliranje kao širi skup. Preciznije, do rezultata se došlo pomoću metoda nadziranog učenja. Model je ostvaren u programskom jeziku R. Rezultati su zadovoljavajući. Predviđanje pobjednika utakmice je uspješno u prosječno 3 od 4 utakmice, a pojedinačni rezultati pripadaju prostoru unutar 5 %. Iako prostora za napredak ima, ovaj rad je pokazao kako kuglanje ipak ima konstantu koja se može iskoristiti za predikcije. Budući rad mogao bi se fokusirati na smanjenje pogreške i odstupanja, testiranje drugih metoda i slično, no metoda nadziranog učenja ispostavila se kao dobra metoda za predviđanje kuglačkih rezultata. The aim of this study was to create a model for predicting the outcomes of bowling matches while gaining a deeper understanding of analyzing the given dataset and handling it effectively. The ultimate goal was to achieve predictions that closely resemble the actual results for both individual and team matches. The chosen approach was predictive modeling, specifically utilizing supervised learning methods and the R programming language as the primary tool. The prediction of match winners was successful in approximately 3 out of 4 matches, and individual scores fell within a 5 % margin. Although there is room for improvement, this study demonstrated that predictive methods can be applied to a sport such as bowling. Future research could focus on reducing errors in individual predictions, exploring alternative methods, and so on. Nonetheless, supervised learning has proven to be an effective approach for predicting 9-pin bowling match results.