Usage of Language Particularities for Semantic Map Construction: Affixes in Russian Language
This paper is devoted to a method of creation the semantic map for Russian language. We consider different approaches for cognitive maps construction which were made for other languages and compare them to the developing algorithm. We also show the main f
- PDF / 186,318 Bytes
- 8 Pages / 439.37 x 666.14 pts Page_size
- 41 Downloads / 137 Views
National Research Nuclear University MEPhI (Moscow Engineering Physics Institute), Moscow, Russian Federation [email protected], [email protected], [email protected], [email protected]
Abstract. This paper is devoted to a method of creation the semantic map for Russian language. We consider different approaches for cognitive maps construc‐ tion which were made for other languages and compare them to the developing algorithm. We also show the main features of Russian words structure and high‐ light the important of them for the further usage in the concept of semantic map. We introduce the set-theoretical model of the Russian words which will be used in further researches. Keywords: Semantic maps · Weak semantic maps · Russian language · Affixes
1
Introduction
In the present time word semantics is given special attention in the Web and in other fields related with automated text recognition. For instance, improving search in the Web (semantic search), where the user’s queries are uniquely interpreted by the searcher, or sentiment analysis (also opinion mining), whose purpose is to determine the attitude of a speaker or a writer with respect to some topic of the overall contextual polarity (e.g. positive or negative) of a document [1]. This concept could be included into many different web-applications, as a rule in social networks, blogs and other resources of personalization, which the semantic search and sentiment analysis are applied in. In any case, automated evaluation of word semantics means using various scales, or dimen‐ sions, that characterize the word meaning [2]. These tools are also known as semantic maps, cognitive maps, semantic spaces. The most popular approaches in this field are based on usage of vector space model. It means that concepts, words or documents (representations) can be associated with vectors in an abstract multidimensional vector space. There are also other approaches the sense of which consists in using manifolds of more complex topology and geometry. In summary, cognitive map can be defined as a topological or metric space, the topology and/or the geometry of which reflect semantic characteristics and relations among a set of representations (such as words or word senses) embedded in this space. Besides, there are other approaches, for example, named “weak semantic cognitive mapping” [3] that is not based on the idea of “dissimilarity”. The idea of the current © Springer International Publishing Switzerland 2016 L. Cheng et al. (Eds.): ISNN 2016, LNCS 9719, pp. 731–738, 2016. DOI: 10.1007/978-3-319-40663-3_84
732
A. Balandina et al.
approach consists in using such notion as “opposite relations” and doesn’t take into account individual semantic characteristics of representations given a priori. Only rela‐ tions, but not semantic features, are given as input. As a result, semantic dimensions of the map that are not predefined to emerge naturally, starting from a randomly generated initial distribution of words in an abstract space with no a priori given semantics and following [3]
Data Loading...
