Using semantic networks in retrieving poems

• Main Author: Sudabeh Nozari
• Format: Article
• Language: fas
• Published: Iranian Research Institute for Information and Technology 2015-12-01
• Series: Iranian Journal of Information Processing & Management
• Subjects: Semantic Networks; Retrieving Poems; Categorization of Hafez’s Sonnets; Allusions (Talmihat) of Hafez
• Online Access: http://jipm.irandoc.ac.ir/browse.php?a_code=A-10-2829-2&slc_lang=en&sid=1

Semantic relationship among existing concepts in a text can be a solution for better and more precise representation of its content. in addition to common indexing methods. Semantic networks are a pace in this way. In this research, we have attempted to extract the concepts related to the Allusions (Talmihat) in Hafez’s sonnets from six most frequently cited commentaries. The implied and literal meanings are extracted and separated subsequently, they are categorized in order to create connection among them and to draw their semantic networks. This effort improves retrieval in Hafez’s poetry based on audience needs. quantitative and qualitative content analysis are employed here. Quantitative content analysis is used to determine the frequency of concepts. Qualitative content analysis is employed to obtain the meanings of verses and to determine the implied and verbal meanings, and explore relationship among concepts, Both has served to draw a semantic network. The concepts and their relationship between them were identified and two semantic networks were drawn. With this contrivance, the distance and proximity of the main node (Hafez) and from each other, and also the number of poems belonging to concept were determined. Thus, tangible understanding of allusions in the poems of Hafez becomes possibleand systematic and, to some extent hierarchical relationship among them can be visualized. Facilitating retrieval is the most important using the semantic networks. Ssearching in these networks: users can 1) find the poems that they need 2) view the position of the subject throughout networks and 3) distinguish relationship among subjects in respect with other diagnoses.