Browsing Doctoral Dissertations by Title
-
Alagić, Mara (Belgrade , 1985)[more][less]
-
Jokanović, Dušan (Podgorica)[more][less]
-
Šandrih, Branislava (Beograd , 2020)[more][less]
Abstract: The main goal of this dissertation is to put different text classification tasks inthe same frame, by mapping the input data into the common vector space of linguisticattributes. Subsequently, several classification problems of great importance for naturallanguage processing are solved by applying the appropriate classification algorithms.The dissertation deals with the problem of validation of bilingual translation pairs, sothat the final goal is to construct a classifier which provides a substitute for human evalu-ation and which decides whether the pair is a proper translation between the appropriatelanguages by means of applying a variety of linguistic information and methods.In dictionaries it is useful to have a sentence that demonstrates use for a particular dictio-nary entry. This task is called the classification of good dictionary examples. In this thesis,a method is developed which automatically estimates whether an example is good or badfor a specific dictionary entry.Two cases of short message classification are also discussed in this dissertation. In thefirst case, classes are the authors of the messages, and the task is to assign each messageto its author from that fixed set. This task is called authorship identification. The otherobserved classification of short messages is called opinion mining, or sentiment analysis.Starting from the assumption that a short message carries a positive or negative attitudeabout a thing, or is purely informative, classes can be: positive, negative and neutral.These tasks are of great importance in the field of natural language processing and theproposed solutions are language-independent, based on machine learning methods: sup-port vector machines, decision trees and gradient boosting. For all of these tasks, ademonstration of the effectiveness of the proposed methods is shown on for the Serbianlanguage. URI: http://hdl.handle.net/123456789/5090 Files in this item: 1
BranislavaSandrihPhd_konacno.pdf ( 9.053Mb ) -
Stojadinović, Tanja (Beograd , 2013)[more][less]
Abstract: Multiplication and comultiplication, which de ne the structure of a Hopf algebra, can naturally be introduced over many classes of combinatorial objects. Among such Hopf algebras are well-known examples of Hopf algebras of posets, permutations, trees, graphs. Many classical combinatorial invariants, such as M obius function of poset, the chromatic polynomial of graphs, the generalized Dehn-Sommerville relations and other, are derived from the corresponding Hopf algebra. Theory of combinatorial Hopf algebras is developed by Aguiar, Bergerone and Sottille in the paper from 2003. The terminal objects in the category of combinatorial Hopf algebras are algebras of quasisymmetric and symmetric functions. These functions appear as generating functions in combinatorics. The subject of study in this thesis is the combinatorial Hopf algebra of hypergraphs and its subalgebras of building sets and clutters. These algebras appear in di erent combinatorial problems, such as colorings of hypergraphs, partitions of simplicial complexes and combinatorics of simple polytopes. The structural connections among these algebras and among their odd subalgebras are derived. By applying the character theory, a method for obtaining interesting numerical identities is presented. The generalized Dehn-Sommerville relations for ag f-vectors of eulerian posets are proven by Bayer and Billera. These relations are de ned in an arbitrary combinatorial Hopf algebra and they determine its odd subalgebra. In this thesis, the generalized Dehn-Sommerville relations for the combinatorial Hopf algebra of hypergraphs are solved. By analogy with Rota's Hopf algebra of posets, the eulerian subalgebra of the Hopf algebra of hypergraphs is de ned. The combinatorial characterization of eulerian hypergraphs, which depends on the nerve of the underlying clutter, is obtained. In this way we obtain a class of solutions of the generalized Dehn-Sommerviller relations for hypergraphs. These results are applied on the Hopf algebra of simplicial complexes. URI: http://hdl.handle.net/123456789/4306 Files in this item: 1
phdTanjaStojadinovic.pdf ( 13.95Mb ) -
Stojadinović, Tanja (Univerzitet u Beogradu , 2014)[more][less]
Abstract: Multiplication and comultiplication, which de ne the structure of a Hopf algebra, can naturally be introduced over many classes of combinatorial objects. Among such Hopf algebras are well-known examples of Hopf algebras of posets, permutations, trees, graphs. Many classical combinatorial invariants, such as M obius function of poset, the chromatic polynomial of graphs, the generalized Dehn-Sommerville relations and other, are derived from the corresponding Hopf algebra. Theory of combinatorial Hopf algebras is developed by Aguiar, Bergerone and Sottille in the paper from 2003. The terminal objects in the category of combinatorial Hopf al- gebras are algebras of quasisymmetric and symmetric functions. These functions appear as generating functions in combinatorics. The subject of study in this thesis is the combinatorial Hopf algebra of hyper- graphs and its subalgebras of building sets and clutters. These algebras appear in di erent combinatorial problems, such as colorings of hypergraphs, partitions of sim- plicial complexes and combinatorics of simple polytopes. The structural connections among these algebras and among their odd subalgebras are derived. By applying the character theory, a method for obtaining interesting numerical identities is pre- sented. The generalized Dehn-Sommerville relations for ag f-vectors of eulerian posets are proven by Bayer and Billera. These relations are de ned in an arbitrary com- binatorial Hopf algebra and they determine its odd subalgebra. In this thesis, the generalized Dehn-Sommerville relations for the combinatorial Hopf algebra of hy- pergraphs are solved. By analogy with Rota's Hopf algebra of posets, the eulerian subalgebra of the Hopf algebra of hypergraphs is de ned. The combinatorial char- acterization of eulerian hypergraphs, which depends on the nerve of the underlying clutter, is obtained. In this way we obtain a class of solutions of the generalized Dehn-Sommerviller relations for hypergraphs. These results are applied on the Hopf algebra of simplicial complexes. URI: http://hdl.handle.net/123456789/3745 Files in this item: 1
phdTanjaStojadinovic.pdf ( 13.95Mb ) -
Berković, Mladen (Beograd , 1977)[more][less]
URI: http://hdl.handle.net/123456789/4100 Files in this item: 1
Konacni_elementi_membrana.PDF ( 3.010Mb ) -
Berković, Mladen (Belgrade , 1977)[more][less]
-
Shafah, Osama (Beograd , 2013)[more][less]
-
Protić, Petar (Novi Sad , 1986)[more][less]
-
Vulanović, Relja (Novi Sad , 1986)[more][less]
-
Šarac, Marica (Belgrade)[more][less]
-
Romano, Daniel (Belgrade , 1985)[more][less]
-
Veljković, Kristina (Beograd , 2016)[more][less]
Abstract: Design of control chart for monitoring central tendency of nongaussian random variables with symmetric or positively skewed distributions is considered. In the case of nongaussian symmetric distributions, modified X bar control chart is proposed in this dissertation. For chosen Student, Laplace, logistic and uniform distributions, theoretical distribution of the standardized sample mean is calculated and approximated with Pearson type II or Pearson type VII distributions. Width of control limits and power of X bar control chart are established, for a given probability of type I error. The results imply that the corresponding Pearson distribution represents very good approximation of the distribution of the standardized sample mean. For implementation of X bar control chart in practice, measures of sample kurtosis are compared and the usage of proposed chart is illustrated on given data. In the case of positively skewed distributions, one sided median control chart for monitoring central tendency of quality characteristics is proposed in this dissertation. For chosen exponential, gamma and Weibull distributions, theoretical distribution of sample median is calculated and approximated with Pearson type I or Pearson type VI distributions. Calculated values of upper control limits and power of median control chart for theoretical distribution of sample median and corresponding Pearson distribution are very close. For implementation of median control chart in practice, measures of sample skewness and sample kurtosis are compared and then proposed median chart is constructed for given data. Besides the statistical design of control charts for monitoring central tendency of nongaussian random variables, their optimal economic statistical design is also considered. Use of genetic algorithms for constrained minimization of expected loss function is proposed in this dissertation. Same symmetric distributions as in the case of statistical design of the X bar control chart and positively skewed distributions as in the case of statistical design of median control chart are chosen. For all chosen distributions of quality characteristic, a corresponding Pearson distribution gives results very close to results based on the theoretical distribution of the standardized sample mean (sample median). URI: http://hdl.handle.net/123456789/4449 Files in this item: 1
Kristinateza.pdf ( 2.651Mb ) -
Zeković, Ana (Beograd , 2015)[more][less]
Abstract: A main focus of the paper is construction of new methods for defining diverse knot distance types - the distance of knots made by crossing changes (Gordian distance) and the distance among knots made by crossing smoothing (smoothing distance). Different ways of knots presentation are introduced, with objective to a mirror curve model. It is presented a purpose of the model, coding of knots, by using the model preferences, as well as introduction of a method to determinate a knots presented by the model and derived all the knots that could be placed to a nets dimensions p×q (p ≤ 4, q ≤ 4). Diverse knot notations are described into details, with a focus to Conway’s notation and its topological characteristics. As it is known, a present algorithms are based on an algebra of chain fractions, that are in close relation with a presentation of rational knots, which results in an absence of a huge number of non-rational knots, in an existing Gordian’s distance tables. The subject of the paper is an implementation of methods with bases on determination of new distances equal 1. The methods are based on a non-minimal presentation of rational and non-rational knots, generation of algorithms established on geometrical characteristics of Conway’s notation and a weighted graph search. The results are organized into Gordian’s distance knots tables up to 9 crossings, and have been enclosed with the paper. In order to append the table with knots having a bigger number of crossings, it has been suggested a method for extension of results for knot families. Using facts of relation among Gordian’s numbers and smoothing numbers, a new method for smoothing number determination is presented, and results in a form of lists for knots not having more then 11 crossings. In conjunction with Conway’s notation concept and the method, algorithms for a smoothing distance are generated. New results are organized in knot tables, up to 9 crossings, combined with previous results, and enclosed with the paper. A changes and smoothing to a knot crossing could be applied for modeling topoisomerase and recombinase actions of DNA chains. It is presented the method for studying changes introduced by the enzymes. A main contribution to the paper is the concept of Conways notation, used for all relevant results and methods, which led to introduction of a method for derivation a new knots in Conways notation by extending C-links. In a lack of an adequat pattern for an existing knot tables in DT-notation, there is usage of a structure based on topological knot concepts. It is proposed a method for knot classification based on Conways notation, tables of all knots with 13 crossings and alternated knots with 14 crossings has been generated and enclosed. The subject of the paper takes into consideration Bernhard-Jablan’s hypothesis for a determination of unknotting number using minimal knot diagrams. The determination is crucial in computation of diverse knot distances. The paper covers one of main problems in knot theory and contains a new method of knot minimization. The method is based on relevance of local and global minimization. 5 There are defined new terms such as a maximum and a mixed unknotting number. The knots that do not change a minimum crossing number, after only one crossing change are taken into consideration for the analyzes. Three classes of the knots are recognized, and called by authors . Kauffman’s knots, Zekovic knots and Taniyama’s knots. The most interesting conclusion correlated with Zekovic knots is that all derived Perko’s knots (for n ≤ 13 crossings) are actually Zekovic knots. Defining this class of knots provides opportunity to emphasize new definitions of specifis featured for well-known Perko’s knots. URI: http://hdl.handle.net/123456789/4255 Files in this item: 1
phdZekovicAna.pdf ( 5.246Mb ) -
Dražić, Milan (Beograd , 1995)[more][less]
-
Dražić, Milan (Beograd , 1995)[more][less]
-
Takači, Arpad (Novi Sad , 1981)[more][less]
-
Mitić, Ljubiša (Belgrade)[more][less]
-
Nikić, Mioljub (Belgrade)[more][less]
-
Demčenko, Vasilije (Belgrade , 1924)[more][less]