IEEE-20180 ICCCNT’12 26 th -28 th July 2012, Coimbatore, India Abstract—Context based Domain identification is a task of identifying correct domain of a word from the sentence. Domain identification is used to resolve ambiguity of words based on context of sentence. Context is a unique proof of meaning and its sentence. Hence, using this unique utility ambiguity is resolved. In this paper, the three stages are discussed which describes the previous and proposed work. This paper briefly focuses on second stage. The second stage describes the technique used to identify domain. After identification how the system is trained to select correct domain is explained. Lastly, the related results are shown. This paper describes usefulness of resolving ambiguity, which helps people to understand meaning of the sentence Keywords- Synonymy; Supervised Training; WordNet Domain; Word Ambiguity. I. INTRODUCTION In many Natural Language Processing applications, the problem of distinguishing between multiple possible senses of word is an important subtask. In order to resolve an ambiguity in a sentence, natural language processing provides word sense disambiguation (WSD) which governs a sentence in which the sense of a word or meaning is used, when the word has multiple meanings (polysemy) [1, 3]. WSD is a process which identifies the correct sense of a word with the help of surrounding words in a sentence. The correct sense of a word is obtained from the context of the sentence. Based on context, a different meaning of the single word in each sentence is associated. Thus, if the word hard work appears near the word sweet, we can say that its meaning related to ‘Success’ and not ‘Taste’ which is known as local context or sentential context[2]. At a onetime Computer that read words, must use a process called word sense disambiguation i.e. finding the correct sense of word [14]. Context may play a role in resolving ambiguity. For example the same piece of information may be ambiguous in one context and unambiguous in another. The context in which an ambiguous word is used often makes it evident which of the meaning is intended. However, some linguistic contexts do not provide sufficient information to disambiguate a used word [5]. To specify the senses a dictionary required for disambiguation process. Then these senses are to be disambiguated. Hence, mass of language data to be disambiguated [12]. For identifying the correct sense of the word the ‘WordNet’ domain is used. A domain may include synsets of different syntactic categories [8]. It groups senses of the same word into homogeneous clusters, with the effect of reducing word polysemy in WordNet. WordNet domain provides semantic domain as a natural way to establish semantic relations among word senses [3, 8]. Domains (such as, for example, biomedical) that meet all these properties are naturally more popular for researchers and developers. Here, we used functionality of the WordNet domain for creating MySQL database named tblFields. This paper describes Word Sense Disambiguation and Domain in section I, section II describes about previous work, section III describes the proposed work and focuses on second stage. In section IV Evaluation of proposed work accuracy is explained. The result of the system is described in section V. Thereafter, section VI describes the conclusion of the system. II. PREVIOUS WORK Under this framework, it has three stages described below: I. In first stage, the database is created to store Fields, General words & Meanings. Also, the POS (Part-of-Speech) Tagger process is implemented to separate the content words [7, 10]. In this, the separation of words is done as step 1 and the target word is picked up from content words as step 2. After, these steps three categories of the words are created as C1, C2 and C3. C1 indicates separated content words, C2 indicates assigned domain of words stored in database and C3 describes maximum count of domain based on context. Based on this the paper had published [17]. Context Based Domain Identification for Resolving Ambiguity Priti Saktel 1 , Urmila Shrawankar 2 IEEE Student Member 2 Department of Computer Science and Engineering G. H. Raisoni College of Engineering Nagpur, MS, INDIA saktel.priti@gmail.com 1 ,urmila@ieee.org 2