<p>U radu je izvr&scaron;eno teoretsko istraživanje i analiza postojećih stavova i re&scaron;enja u oblasti validacije i provere kvaliteta modela podataka. Kreiran je teorijski model ontolo&scaron;ki zasnovane analize semantičke korektnosti modela podataka primenom sistema automatskog rezonovanja i izvr&scaron;ena praktična implementacija teorijskog modela, &scaron;to je potvrđeno i sprovedenim eksperimentalnim istraživanjem. Razvijena je softverska aplikacija za formalizaciju modela podataka i mapiranje ontologije u oblik Prolog klauzula. Formirana su pravila zaključivanja na predikatskom računu prvog reda, koja su integrisana sa modelom podataka i domenskom ontologijom. Upitima u okviru Prolog sistema, vr&scaron;i se provera semantičke korektnosti modela podataka. Definisana je i metrika ontolo&scaron;kog kvaliteta modela podataka koja se bazira na odgovorima sistema automatskog rezonovanja.</p> / <p>Work presents a theoretical study and analysis of existing theories and solutions in the area of data model validation and quality checking. It is created a theoretical model of ontology based analysis of data model semantic correctness by applying automated reasoning system which is practicaly implemented and confirmed by the conducted experimental research. A software application is developed for data model formalization and ontology mapping in Prolog clauses form. Reasoning rules are formed the in first-order predicate logic, which are integrated with the data model and domain ontology. Semantic correctness of the data model is checked with queries within Prolog system. Metrics of ontological quality of the data model are defined which are based on automated reasoning system replies.</p>

<p>Doktorska disertacija pripada oblasti<br />informacionih sistema, odnosno podoblasti koja<br />se bavi upravljanjem skladi&scaron;tenjem i<br />pretraživanjem informacija. Osnovni cilj<br />disertacije je modeliranje i implementacija<br />skupa alata koji omogućavaju upotrebu fazi<br />logike u radu sa relacionim bazama podataka.<br />Da bi se do tog skupa alata do&scaron;lo, najpre je<br />relacioni model podataka pro&scaron;iren elementima<br />teorije fazi skupova, a zatim je definisano fazi<br />pro&scaron;irenje upitnog jezika SQL &ndash; PFSQL.<br />Interpreter za taj jezik je implementiran u<br />okviru fazi JDBC drajvera koji, osim<br />implementacije interpretera, sadrži i elemente<br />koji omogućavaju jednostavnu upotrebu ovih<br />mehanizama iz programskog jezika Java. Skup<br />alata je zaokružen implementacijom CASE<br />alata za razvoj fazi-relacionog modela baze<br />podataka. Osim toga, razmatrane su i<br />mogućnosti za upotrebu PFSQL jezika u<br />vi&scaron;eslojnim aplikacijama.</p> / <p>This doctoral dissertation belongs to the<br />field of information systems, subfield<br />information storage and retrieval management.<br />The main subject of the dissertation is modeling<br />and implementation of a set of tools that allow<br />usage of fuzzy logic in relational database<br />applications<br />In order to achieve that goal, at first, the<br />relational data model is extended with elements<br />of fuzzy set theory. After that, a fuzzy<br />extension of the SQL query language, called<br />PFSQL, is defined. An interpreter for that<br />language is implemented as a part of the fuzzy<br />JDBC driver. Beside the implementation of the<br />interpreter, this fuzzy JDBC driver contains<br />elements that allow simple usage of offered<br />mechanisms from Java programming language.<br />The set of tools is concluded with the<br />implementation of the CASE tool for the<br />development of fuzzy-relational data models. In<br />addition, possibilities to use PFSQL language<br />on the middle tier of multi tier systems are<br />discussed.</p>

<p>Rad je posvećen problemu rane i brze detekcije i izolacije neželjenih stanja dinamičkih sistema, sa posebnim naglaskom na rano otkrivanje različitih nepravilnosti u radu i kvarova industrijskih procesa.</p> / <p>The thesys is dedicated to the problem of early and swift detection and isolation of unwanted working regimes of dynamical systems, with particular emphasis on the early detection of various irregularities and failures of industrial processes.</p>

<p>Doktorska disertacija se bavi temama rešavanja računarski teških<br />problema kombinatorne optimizacije. Istaknut je problem maksimalne<br />klike kao predstavnik određenih struktura u grafovima. Problem<br />maksimalne klike i sa njim povezani problemi su formulisani kao<br />nelinearne funkcije. Rešavani su sa ciljem otkrivanja novih metoda<br />koje pronalaze dobre aproksimacije rešenja za neko razumno vreme.<br />Predložene su varijante Metode promenljivih okolina na rešavanje<br />maksimalne klike u grafu. Povezani problemi na grafovima se mogu<br />primeniti na pretragu informacija, raspoređivanje, procesiranje<br />signala, teoriju klasifikacije, teoriju kodiranja, itd. Svi algoritmi<br />su implementirani i uspešno testirani na brojnim različitim<br />primerima.</p> / <p>This Ph.D. thesis addresses topics NP hard problem solving approaches in<br />combinatorial optimization and according to that it is highlighted maximum<br />clique problem as a representative of certain structures in graphs. Maximum<br />clique problem and related problems with this have been formulated as non<br />linear functions which have been solved to research for new methods and<br />good solution approximations for some reasonable time. It has been<br />proposed several different extensions of Variable Neighborhood Search<br />method. Related problems on graphs could be applied on information<br />retrieval, scheduling, signal processing, theory of classi_cation, theory of<br />coding, etc. Algorithms are implemented and successfully tested on various<br />different tasks.</p>

<p>Ispitano je prisustvo toksičnih cijanobakterija u različitim ekosistemima sa teritorije&nbsp;Republike Srbije i analizirane su negativne posledice koje mogu da prouzrokuju ovi&nbsp;mikroorganizmi. Formirana je Baza podataka cijanobakterija u Srbiji koja prilaže veliki broj bitnih&nbsp; i korisnih informacija iz preko 70 literaturnih izvora o prostiranju i učestalosti pojave cijanobakterija i njihovih toksina u periodu od 130 godina, kao i njihovih efekata na živi svet u vodenim ekosistemima, ali i &scaron;ire. Istraživana su 64 vodena ekosistema, uključujući reke, jezera, bare, kanale, ribnjake, akumulacije za navodnjavanje, akumulacije za snabdevanje vodom za piće i akumulacije sa drugom namenom, gde je najče&scaron;će cvetalo pet vrsta cijanobakterija i to:&nbsp;<em> Microcystis aeruginosa,&nbsp; Aphanizomenon flos-aquae</em>,&nbsp; <em>Planktothrix agardhii,&nbsp; Microcystis flos-aquae</em>&nbsp; i<em>&nbsp; Planktothrix rubescens</em>&nbsp; koje ujedno i &scaron;ire svoj areal rasprostranjenja na teritoriji Republike Srbije. U brojnim vodenim telima detektovani su mikrocistini i to u visokim koncentracijama.</p><p>Istraživanje vodenog ekosistema Ludo&scaron; sa teritorije Republike Srbije vr&scaron;eno je da bi se ustanovilo prisustvo i uticaj cijanobakterija i cijanotoksina na druge biljne i životinjske organizme u prirodnim uslovima. Trofički status jezera Ludo&scaron; najče&scaron;će je eutrofan, a cvetanje cijanobakterija kontinuirano jo&scaron; od 1970. godine. Tokom 2011. i 2012. godine&nbsp; vrste<em> Limnothrix&nbsp; redekei&nbsp;</em> i&nbsp;<em> Pseudanabaena&nbsp; limnetica</em>&nbsp; nađene&nbsp; u&nbsp; cvetanju. Detektovano je i prisustvo mikrocistina/nodularina i saksitoksina u biomasi i vodi, a mikrocistini su detektovani i u tkivu vodenih biljaka (trske&nbsp;<em> Phragmites communis</em>, rogoza&nbsp; <em>Typha latifolia</em>&nbsp; i ljubičastog lokvanja<em> Nymphaea elegans</em>) i ribi (<em>Carassius gibelio</em>) iz jezera Ludo&scaron;. Histolo&scaron;kim pregledom tkiva ribe&nbsp; pronađene su naizraženije promene u jetri, bubrezima i &scaron;krgama, a primećene su i na crevima.</p><p>Testiranjem biolo&scaron;kih lesnih pokorica sa teritorije Vojvodine nisu detektovani&nbsp;mikrocistini/nodularini, a nije detektovana ni toksičnost uzoraka. Pretpostavlja se da su koncentracije cijanotoksina ispod granica detekcije ili ih nema u testiranim biolo&scaron;kim lesnim pokoricama. Razvojem novih metoda i optimizacijom postojećih za detekciju cijanotoksina u biolo&scaron;kim lesnim pokoricama i drugim terestričnim ekosistemima potrebno je proveriti dobijene rezultate.</p><p>Proučavanjem svojstva 84 soja cijanobakterija iz NSCCC, koji potiču iz terestričnih i&nbsp;vodenih ekosistema sa teritorije Republike Srbije, dobijena je intracelularna toksičnost u eksponencijalnoj i stacionarnoj fazi rasta, kao i ekstracelularna toksičnost &nbsp;kod jednog soja koji potiče sa terestričnog ekosistema. Dobijeni su i pozitivni rezultati &nbsp;na prisustvo mikrocistina, nodularina ili/i saksitoksina kod 34,1% terestričnih i 55,5% vodenih sojeva koji potiču sa teritorije Republike Srbije. Rezultati ukazuju na&nbsp; potencijalnu opasnost pojave ovih&nbsp; mikroorganizama i njihovih toksičnih metabolita u različitim ekosistemima sa teritorije Republike Srbije.</p><p>Eksperimentalno je potvrđena akumulacija mikrocistina u račiću<em>&nbsp; Daphnia pulex&nbsp; </em>usled ishrane toksičnim sojem iz NSCCC, zbog čega upotreba ove vrste račića u izvođenju biolo&scaron;kih testova za testiranje prisustva cijanotoksina treba da se preispita.&nbsp; &Scaron;tavi&scaron;e, veliki broj jedinki dafnija (92,2%) koristio je istraživane sojeve iz NSCCC za ishranu, odnosno ishrana je bila moguća sa svim vodenim i sa gotovo 90% terestričnih sojeva, &scaron;to može poslužiti kao osnov za dalja istraživanja prevencije cvetanja.</p><p>S obzirom na mogućnost ishrane račića&nbsp;<em> Daphnia&nbsp; sp</em>. cijanobakterijama, kompleks&nbsp;<br />ribnjaka sa teritorije Republike Srbije kori&scaron;ćen je za istraživanje potencijalnog načina&nbsp;<br />prevencije pojave i cvetanja cijanobakterija u zavisnosti od pravovremenog uno&scaron;enja&nbsp;<br />pomenutog račića. Na osnovu koncentracije hlorofila&nbsp; a&nbsp; i trofičkog statusa,&nbsp; kvalitativne i kvantitativne analize cijanobakterija, toksičnosti vode&nbsp; i prisustva cijanotoksina mikrocistina/nodularina i saksitoksina, potvrđeno je smanjenje cvetanja cijanobakterija i drugih negativnih efekata u eksperimentalnim jezerima u odnosu na kontrolna. Ukoliko pak dođe do cvetanja i proizvodnje toksina&nbsp; ijanobakterija u vodenim ekosistemima, zbog ozbiljnost pojave i mogućih negativnih posledica po zdravlje ljudi, neohodno je uvesti postupke eliminacije ćelija&nbsp; cijanobakterija i njihovih &nbsp;toksina u praksu pri obradi&nbsp; otpadnih voda i preči&scaron;ćavanja vode iz povr&scaron;inskih akumulacija u Republici Srbiji.</p> / <p>The presence of toxic&nbsp; cyanobacteria in different ecosystems from the territory of the Republic of Serbia was analyzed as well as the negative consequences that may be causeed by these microorganisms. Serbian Cyanobacterial Data Base was formed where great number of important and useful information from over 70 literary sources regarding the distribution and frequency of cyanobacteria and their toxins over a period of 130 years, as well as their effects on wildlife in aquatic ecosystems, and beyond was presented. The study consisted of 64 aquatic ecosystems, including rivers, lakes, ponds, canals, irrigation reservoirs, reservoirs for drinking water supply and reservoirs with other purpose, where five species of cyanobacteria <em>Microcystis aeruginosa,&nbsp; Aphanizomenon flos-aquae,&nbsp; Planktothrix agardhii,&nbsp; Microcystis flosaquae&nbsp;</em> and&nbsp; <em>Planktothrix rubescens</em>&nbsp; frequently bloomed, and also expanded their area of distribution on the territory of the Republic of Serbia. In many water bodies microcystins were detected in high concentrations.</p><p>Research into aquatic ecosystem Ludo&scaron;, located on the territory of the Republic of Serbia, was performed in order to determine the presence and effect of cyanobacteria and cyanotoxins on other plant and animal organisms in natural conditions. Trophic status of&nbsp; the lake Ludo&scaron; was usually eutrophic, and cyanobacteria bloom is constant since 1970. During 2011 and 2012, the species <em>Limnothrix&nbsp; redekei</em>&nbsp; and&nbsp; <em>Pseudanabaena&nbsp; limnetica&nbsp;</em> were found in blooming. Presence of microcystins/nodularin and saxitoxin was detected in&nbsp; biomass and water, and microcystins were detected in tissues of aquatic plants (reed <em>Phragmites communis,</em> cattail&nbsp;<em>Typha latifolia&nbsp;</em>and royalblue waterlily&nbsp; <em>Nymphaea elegans</em>) and fish (<em>Carassius gibelio</em>) from the lake Ludo&scaron;. Histological examination of tissue showed most prominent changes in liver, kidney and gills, and alterations were also observed in the intestine.</p><p>Testing of the biological loess crust from Vojvodina showed no presence of microcystins/nodularin and toxicity of samples was not detected as well. It is assumed that cyanotoxin concentrations were below detection limit or are absent from the tested biological loess crusts. The development of new methods and optimization of existing ones for detection of cyanotoxins in biological loess crustsand&nbsp; other terrestrial ecosystems is necessary in order to revise obtained results.</p><p>Research of the properties from 84 strains of cyanobacteria from NSCCC originating&nbsp;from terrestrial and aquatic ecosystems from the territory of the Republic of &nbsp;Serbia, resulte d inintracellular toxicity in exponential and stationary growth phase, as &nbsp;well as extracellular toxicity of a strain originating from terrestrial ecosystems. The obtained results were positive for the presence of microcystins, nodularin and/or saxitoxin&nbsp; for 34.1% terrestrial and 55.5% aquatic strains originating from the territory of the Republic of Serbia. These results demonstrate the potential risk of occurrence of these microorganisms and their metabolites in different ecosystems from the territory of the Republic of Serbia.</p><p>Accumulation of microcystins in shrimp <em>Daphnia pulex</em>&nbsp;after feeding with toxic strain&nbsp;from NSCCC was experimentally confirmed, indicating that the use of this species of shrimp in biological tests which determine the presence of cyanotoxins needs to be revisited. Moreover, a large number of&nbsp;<em>Daphnia&nbsp; individuals </em>(92.2%) used the investigated strains of NSCCC for food, and the feeding was possible with all &nbsp;the water strains and nearly 90% of terrestrial strains, which can serve as a basis for further research of bloom prevention.</p><p>With regard to the possibility shrimp <em>Daphnia&nbsp;</em>sp. feeding with cyanobacteria, a&nbsp;complex of ponds from the territory of the Republic of Serbia was used to explore &nbsp;potential ways of prevention the occurrence and blooming of cyanobacteria, depending on the timely introduction of the aforementioned shrimp. Based on the concentrations of chlorophyll&nbsp; a&nbsp; and trophic status, qualitative and quantitative analysis of cyanobacteria, the toxicity of water and the presence of&nbsp; cyanotoxins microcystins/nodularin and saxitoxin, reduction in blooming cyanobacteria and other negative effects in the experimental lakes when compared to the control lakes was confirmed. If the blooming regardless occurres as well as production of cyanobacterial &nbsp;toxins in aquatic ecosystems, due to the seriousness of the phenomenon and the possible negative consequences for human health, it would be necessary to introduce &nbsp;procedures for the elimination of cells of cyanobacteria and their toxins into practice in waste water treatment and purification of water from surface reservoirs in the Republic of Serbia.</p>

<p>Cilj &ndash; Cilj istraživanja u okviru doktorske disertacije je razvoj sistema za automatsku polu-nadgledanu klasifikaciju podataka. Sistem bi trebao biti primenljiv na &scaron;irokom spektru domena gde je neophodna klasifikacija podataka, a te&scaron;ko je, ili čak nemoguće, doći do dovoljno velikog i raznovrsnog obučavajućeg skupa podataka<br />Metodologija &ndash; Modeli opisani u disertaciji se baziraju na kombinaciji ko-trening algoritma i tehnika učenja sa grupom hipoteza. Prvi korak jeste obučavanje grupe klasifikatora velike raznolikosti i kvaliteta. Sa ovim ciljem modeli eksploati&scaron;u primenu različitih konfiguracija ko-trening algoritma na isti skup podataka. Prednost ovog pristupa je mogućnost kori&scaron;ćenja značajno manjeg anotiranog obučavajućeg skupa za inicijalizaciju algoritma.<br />Skup nezavisno obučenih ko-trening klasifikatora se kreira generisanjem predefinisanog broja slučajnih podela obeležja polaznog skupa podataka. Nakon toga se, polazeći od istog inicijalnog obučavajućeg skupa, ali kori&scaron;ćenjem različitih kreiranih podela obeležja, obučava grupa ko-trening klasifikatora. Nakon ovoga, neophodno je kombinovati predikcije nezavisno obučenih klasifikatora.<br />Predviđena su dva načina kombinovanja predikcija. Prvi način se zasniva na klasifikaciji zapisa na osnovu većine glasova grupe ko-trening klasifikatora. Na ovaj način se daje predikcija za svaki od zapisa koji su pripadali grupi neanotiranih primera kori&scaron;ćenih u toku obuke ko-treninga. Potom se primenjuje genetski algoritam u svrhu selekcije najpouzdanije klasifikovanih zapisa ovog skupa. Konačno,<br />163<br />najpouzdanije klasifikovani zapisi se koriste za obuku finalnog klasifikatora. Ovaj finalni klasifikator se koristi za predikciju klase zapisa koje je neophodno klasifikovati. Opisani algoritam je nazvan Algoritam Statistike Slučajnih Podela (Random Split Statistics algorithm, RSSalg).<br />Drugi način kombinovanja nezavisno obučenih ko-trening klasifikatora se zasniva na GMM-MAPML tehnici estimacije tačnih klasnih obeležja na osnovu vi&scaron;estrukih obeležja pripisanih od strane različitih anotatora nepoznatog kvaliteta. U ovom algoritmu, nazvanom Integracija Vi&scaron;estrukih Ko-treninranih Klasifikatora (Integration of Multiple Co-trained Classifiers, IMCC), svaki od nezavisno treniranih ko-trening klasifikatora daje predikciju klase za svaki od zapisa koji je neophodno klasifikovati. U ovoj postavci se svaki od ko-trening klasifikatora tretira kao jedan od anotatora čiji je kvalitet nepoznat, a svakom zapisu, za koga je neophodno odrediti klasno obeležje, se dodeljuje vi&scaron;e klasnih obeležja. Na kraju se primenjuje GMM-MAPML tehnika, kako bi se na osnovu dodeljenih vi&scaron;estrukih klasnih obeležja za svaki od zapisa izvr&scaron;ila estimacija stvarnog klasnog obeležja zapisa.<br />Rezultati &ndash; U disertaciji su razvijena dva modela, Integracija Vi&scaron;estrukih Ko-treninranih Klasifikatora (IMCC) i Algoritam Statistike Slučajnih Podela (RSSalg), bazirana na ko-trening algoritmu, koja re&scaron;avaju zadatak automatske klasifikacije u slučaju nepostojanja dovoljno velikog anotiranog korpusa za obuku. Modeli predstavljeni u disertaciji dizajnirani su tako da omogućavaju primenu ko-trening algoritma na skupove podataka bez prirodne podele obeležja, kao i da unaprede njegove performanse. Modeli su na vi&scaron;e skupova podataka različite veličine, dimenzionalnosti i redudantnosti poređeni sa postojećim ko-trening alternativama. Pokazano je da razvijeni modeli na testiranim skupovima podataka postižu bolje performanse od testiranih ko-trening alternativa.<br />Praktična primena &ndash; Razvijeni modeli imaju &scaron;iroku mogućnost primene u svim domenima gde je neophodna klasifikacija podataka, a anotiranje podataka dugotrajno i skupo. U disertaciji je prikazana i primena razvijenih modela u nekoliko konkretnih<br />164<br />situacija gde su modeli od posebne koristi: detekcija subjektivnosti, vi&scaron;e-kategorijska klasifikacija i sistemi za davanje preporuka.<br />Vrednost &ndash; Razvijeni modeli su korisni u &scaron;irokom spektru domena gde je neophodna klasifikacija podataka, a anotiranje podataka dugotrajno i skupo. Njihovom primenom se u značajnoj meri smanjuje ljudski rad neophodan za anotiranje velikih skupova podataka. Pokazano je da performanse razvijenih modela prevazilaze performanse postojećih alternativa razvijenih sa istim ciljem relaksacije problema dugotrajne i mukotrpne anotacije velikih skupova podataka.</p> / <p>Aim &ndash; The research presented in this thesis is aimed towards the development of the system for automatic semi-supervised classification. The system is designed to be applicable on the broad spectrum of practical domains where automatic classification of data is needed but it is hard or impossible to obtain a large enough training set.<br />Methodology &ndash; The described models combine co-training algorithm with ensemble learning with the aim to overcome the problem of co-training application on the datasets without the natural feature split. The first step is to create the ensemble of co-training classifiers. For this purpose the models presented in this thesis apply different configurations of co-training on the same training set. Compared to existing similar approaches, this approach requires a significantly smaller initial training set.<br />The ensemble of independently trained co-training classifiers is created by generating a predefined number of random feature splits of the initial training set. Using the same initial training set, but different feature splits, a group of co-training classifiers is trained. The two models differ in the way the predictions of different co-training classifiers are combined.<br />The first approach is based on majority voting: each instance recorded in the enlarged training sets resulting from co-training application is classified by majority voting of the group of obtained co-training classifiers. After this, the genetic algorithm is applied in order to select the group of most reliably classified instances from this set. The most reliable instances are used in<br />167<br />order to train a final classifier which is used to classify new instances. The described algorithm is called Random Split Statistic Algorithm (RSSalg).<br />The other approach of combining single predictions of the group of co-training classifiers is based on GMM-MAPML technique of estimating the true hidden label based on the multiple labels assigned by multiple annotators of unknown quality. In this model, called the Integration of Multiple Co-trained Classifiers (IMCC), each of the independently trained co-training classifiers predicts the label for each test instance. Each co-training classifier is treated as one of the annotators of unknown quality and each test instance is assigned multiple labels (one by each of the classifiers). Finally, GMM-MAPML technique is applied in order to estimate the true hidden label in the multi-annotator setting.<br />Results &ndash; In the dissertation the two models are developed: the Integration of Multiple Co-trained Classifiers (IMCC) and Random Split Statistic Algorithm (RSSalg). The models are based on co-training and aimed towards enabling automatic classification in the cases where the existing training set is insufficient for training a quality classification model. The models are designed to enable the application of co-training algorithm on datasets that lack the natural feature split needed for its application, as well as with the goal to improve co-training performance. The models are compared to their co-training alternatives on multiple datasets of different size, dimensionality and feature redundancy. It is shown that the developed models exhibit superior performance compared to considered co-training alternatives.<br />Practical application &ndash; The developed models are applicable on the wide spectrum of domains where there is a need for automatic classification and training data is insufficient. The dissertation presents the successful application of models in several concrete situations where they are highly<br />168<br />beneficial: subjectivity detection, multicategory classification and recommender systems.<br />Value &ndash; The models can greatly reduce the human effort needed for long and tedious annotation of large datasets. The conducted experiments show that the developed models are superior to considered alternatives.</p>

No description available.

<p>Za potrebe evaluacije kvaliteta vizuelizacije u svrhu prakse inkluzivnog dizajna razvijen je namenski jezik koji hendikepe vida opisuje kroz formalizam baziran na vizuelnim promenljivama. Namenski jezik se može koristiti kao dokumentacija, ili kao specifikacija hendikepa za posebno napisan simulator vizuelnih hendikepa koji se može koristiti za potrebe testiranja. Kao pandan ovoj tehnologiji razvijen je novi mehanizam za evaluaciju vizuelizacija koji kroz inovacije u metodologiji i statistici omogućava da se merodavni zaključci izvedu iz relativno malih uzoraka podataka.</p> / <p>For purposes of visualization quality evaluation as part of the practice of<br />inclusive design, a domain specific language was developed such that it<br />describes visual handicaps through a formalism based on visual variables.<br />This domain specific language can be used as documentation or as a handicap<br />specification for a specially constructed visual handicap simulator usable for<br />testing. Alongside this technology, a novel mechanism for visualization<br />evalution was developed. This mechanism allows, through innovations in<br />metholodoly and statistics, the production of trustworthy results from relatively<br />small data sets.</p>

<p>Prevencija u osiguranju motornih vozila postaje permanentan proces, koji se usavršava i nadograđuje. Povećan rizik prevara u osiguranju zahteva automatizaciju procesa osiguranja i procene šteta na motornim vozilima, kao i modelovanje samog procesa uz određen nivo razmene podataka. Funkcionisanje celokupnog sistema zavisi od definisanja elemenata koji dovode do povećanog rizika prevara u osiguranju motornih vozila. Uspostavljanje procesa automatizacije će takve rizike umanjiti i ublažiti a ujedno povećati i obim osiguranja. Na bazi dobijenih podataka će biti omogućeno filtriranje i blagovremena automatska selekcija a sistem primenjiv u osiguranju.</p> / <p>Prevention in motor insurance becomes a permanent process, which is<br />improved and upgraded. Increased risk of fraud in insurance claims process<br />automation and security assessments of damage to motor vehicles, as well<br />as the modeling of the process with a certain level of data exchange. The<br />functioning of the entire system depends on the definition of the elements<br />that lead to an increased risk of fraud in motor insurance. Establishment of<br />process automation will reduce such risks and also increase insurance<br />coverage. Based on the obtained data will be able to filter and timely<br />automatic selection and the system will be able to be effective in insurance.</p>

<p>Doktorskom disertacijom je dokazano da postojeće metode detekcije i identifikacije loših podataka nisu primenjive na distributivne mreže usled njihovih specifičnosti u stepenu redundanse merenja i broja pseudo merenja. Dodatno, razvijeni su algoritmi detekcije loših oblasti primenom dekuplovanog Hi-kvadrat testa, identifikacije loših merenja primenom novo definisanih izbeljenih reziduala, estimacije fazne konektivnosti primenom uslovnih ograničenja u estimatoru stanja, i korekcije pseudo merenja primenom informacija sa pametnih brojila. Navedeni algoritmi su specijalizovani za distributivne mreže i verifikovani primenom na dva test sistema.</p> / <p>The doctoral dissertation has demonstrated that conventional bad data detection and<br />identification methods cannot be efficiently applied in distribution networks, due to<br />their characteristics such as low measurement redundancy, number of pseudo<br />measurements and level of measurements correlation. In addition, the doctoral<br />dissertation described newly developed algorithms for bad area detection based on<br />decoupled Chi-squares test, bad data identification using newly defined whitened<br />residuals, estimation of phase connectivity by extension of state estimation with<br />conditional constraints and correction of pseudo measurements using AMI data. The<br />mentioned algorithms are specialized for distribution networks and verified through<br />simulation on two test systems.</p>