Application of the NOK method in sentence modelling M. Rauker Koch*, M. Pavli**, A. Jakupovi*** *Railway and technical school Moravice, Vrbovsko, Croatia ** Department of Informatics, University of Rijeka *** Business Department, Polytechnic of Rijeka marina.rauker@skole.hr , mile.pavlic@ris.hr , alen.jakupovic@veleri.hr Abstract - Knowledge representation is one of the areas covered by artificial intelligence. One of the methods for graphical representation of text expressed knowledge is the method NOK (Nodes Of Knowledge). NOK method enables transformation of text expressed knowledge into a graphical network of words and group of words. In this paper application of NOK method is presented. This application is based on sentences from an Aesop's Fable in Croatian ("Golden eggs in the chicken") and English ("The Goose with the golden eggs") version. In this way the applicability of this method on two natural human languages is presented, and similarities and differences that are partially conditioned by freedom of translators, and not only by differences in the syntax of the two languages, are observed. Key words: Node Of Knowledge, Text-based knowledge representation, Network-based knowledge representation. I. INTRODUCTION AI studies computational modelling of human intelligence. Intelligent behaviour is conditioned by knowledge – human act based on what they know (or believe) about the world [1], [2]. One of the fields of interest of AI is knowledge representation and reasoning, which is suitable for reaching artificially intelligent behaviour. Many methods of converting knowledge into a form that can be encoded on a computer have been developed and they are applicable to different types of knowledge [3], [4]. Graphical method Nodes Of Knowledge designed for the presentation of sentences written in natural language is applied in this paper [5]. The formalization of this method is shown in [6]. The possibility of its application has been demonstrated on one text in Croatian and English. By analyzing produced graphic presentations answers to the following questions were obtained: Is there a difference in the complexity of making diagrams, which model is more complex, can answers to the same questions be read from both diagrams, regardless of differences in translation, are there any questions applicable on one model only and how can this problem be solved. II. REVIEW OF RELEVANT PUBLICATION In the field of artificial intelligence different approaches to knowledge representation formalisms are defined [2], [3], [7]. Conceptual graphs [8], [9], [10], semantic networks [11], frames [12] and predicate logic [13] are traditional formalisms that eventually produced numerous other formalisms and methods for knowledge representation. Ontologies have had a significant impact in the field of knowledge representation in the past twenty years [14], [15], [16]. They emphasize the semantic components of knowledge and language [17]. Furthermore, the field of computer analysis of natural language (NLP) is also involved in the formalization of knowledge expressed in language and text [18], [19], [20], [21], [22], [23], [24], [25], [26], [27]. There are two approaches to the formalisation of text- expressed knowledge: declarative and procedural. Examples of declarative approaches are: BG (Basic Conceptual Graphs) [28], MULTINET (Multi-layered extended semantic networks) [29], HSF (Hierarchical Semantic Form) [30], RDF (Resource Description Framework) [31], and other methods based on the first order predicate calculus (FOPC). The result of the procedural approach is a rule system. III. RESEARCH MOTIVATION Text as a knowledge carrier is not suitable for computer-supported knowledge dynamics (changes, paraphrasing, distribution, use, generation of new knowledge, abandonment, etc.). It should be specially formatted - formalised - in order to enable a computer to perform mathematical/logical operations. Nodes of Knowledge (NOK) method is a new method by which sentences written in natural language can be modelled. The purpose of graphical modelling sentences with NOK method is the creation of models (formal representation) of different forms of sentences (declarative, interrogatory, simple, complex, etc.) in order to verify the basic concepts of the method. Designed models and their verification represent the basis for the development of mathematical formalism for knowledge representation and reasoning based on a NOK method - first formalism is shown in [6]. The aim of this paper is to model the same text in two languages, English and Croatian, compare derived models and notice similarities and differences in the model that are conditioned by the freedom of translators. IV. NODES OF KNOWLEDGE NOK method belongs to a group of semantic networks [11] where knowledge is represented as a graph [32]. Its 1426 MIPRO 2014/CIS