ISSN : 2319-7323

INTERNATIONAL JOURNAL OF COMPUTER SCIENCE ENGINEERING

ABSTRACT

Title	:	SpatioTag: A Strategic Approach for Recommending Image Tags for Socially Relevant Photos using Spatial Information
Authors	:	Pushpa C N, Pradeepa C, Gerard Deepak, Thriveni J, Venugopal K R
Keywords	:	eLearning; MOOCs; DTBP;eLearning adoption;eLearning behaviour; Education
Issue Date	:	Nov-Dec 2018
Abstract	:	Social Networking is one of the most successful trends to share friendships over the World Wide Web and it has emerged as a repository of a large number of photographs as social actors share their photographs over the Social networking platform. Every photograph uploaded on the Social networking site requires social tagging to structure and organize the photographs that are posted by individuals. In order to achieve, the best quality of social tags without any falsified or over ridden tag information, a tag recommendation strategy by automatically assigning relevant tags is implemented with a focus given to user preferences. The user defined geographical location is considered to geographically locate the photo. Tags that are semantically similar for a specific photo are analyzed and the best tag is ordered first. In case of several heterogeneous tags availability for a photograph, a methodology to identify the most relevant tag by ordering them to their context is proposed. The System is specific to Community Contributed Photos and thus to ensure the tag validity and increase the elements in recommended tags, content based similarity and text based similarity is computed using the Normalized Pointwise Mutual Information Strategy. Experiments were conducted on a large data sets collected from Flickr and further integrated into the social tagger for verification of the efficiency of the proposed system. An overall accuracy percentage of 82.02 is achieved by the proposed SpatioTag System.
Page(s)	:	261-267
ISSN	:	2319-7323
Source	:	Vol. 7, No. 7