International Journal of Computer Applications (0975 – 8887) Volume 129 – No.5, November2015 16 Comparative Study of Semantic Search Techniques using RDF Salil C. Damle Department of Computer Engineering Dwarkadas J. Sanghvi College of Engineering Mumbai, India Shreyas S. Kupekar Department of Computer Engineering Dwarkadas J. Sanghvi College of Engineering Mumbai, India Khushali Deulkar Assistant Professor Department of Computer Engineering Dwarkadas J. Sanghvi College of Engineering Mumbai, India ABSTRACT In the sea of information available to us today, the prevalent searching techniques using keywords and ranking algorithms fall short on many aspects. In such a scenario, the emergence of semantic searching techniques is attempting to fill this void. It is required by Semantic searching techniques to understand the intent of users and the meaning of the query entered using Natural Language. The query is semantically broken down and stored in the Resource Description Framework (RDF) format. Such a structured format is more useful for implementing the search. In this paper, a comparison of various searching techniques which make use of RDF is done, based on certain parameters. The main intention is to identify the searching techniques which will be best suited for different purposes. General Terms Semantic Search, Algorithms. Keywords RDF, RDF(s), Semantic Web, ontology. 1. INTRODUCTION Today, the web is evolving from a simple keyword extraction based technique to a more advanced and useful semantic way. The intention has now changed to understanding the reasons for which a user posts a query rather than to simply search for articles and documents which have the words from the query. Semantic web [1] thus forms the next generation of internet i.e. web 3.0. It provides a standard data format and exchange protocols which delivers fast results and at the same time more relevant results to the user. From the early web solutions of HTML which could simply illustrate documents and provide linking between them, recent advancements have now come to semantic HTML [2], Extensible Markup Language (XML), Web Ontology Language (OWL) and Resource Description Framework (RDF) which can describe arbitrary things such as objects, people, places etc. in a structured way. The Resource Description Framework [3] is a World Wide Web Consortium (W3C) specification which is used to convert a natural language query into a tabular format that can be easily shared and used universally. It was designed as an improved data model over the likes of the class diagram or the E-R model. The RDF constructs a Metadata format which is used to store a query or descriptions of web resources in a subject-predicate-object configuration also known as the triples format in the RDF nomenclature [4]. The simple directed multi-graph format and the capacity of modelling abstract information has made it an essential component for usage in Knowledge management applications. This paper has been organised into various sections as follows: Section 2 introduces the parameters which have been used to compare the approaches. Section 3 summarizes and talks about the papers which have been used for the comparative analysis. The next section i.e. the section 4 compares the papers selected in the Section 3 on the basis of the parameters described in the section 2. The comparison observations are concluded in the section 5. Sections 6 and 7 have the acknowledgements and references respectively. 2. CLASSIFICATION PARAMETERS This section mentions and explains the parameters on the basis of which the comparing of RDF based semantic searching techniques has been conducted. The criteria are so decided because they are best able to provide a clear idea about the approaches used and to supply a distinct classification between them. The classification parameters used in this paper are as follows: Methodology and Indexing, Evaluation of performance, Advantages and Disadvantages. There are other parameters to classify the searching techniques which have not been discussed in this survey. Some of them include Ranking model, inferencing, and ontology construction. They haven‟t been included as the parameters chosen for the comparison are sufficient to classify the search techniques and some parameters are not clearly identifiable through the papers. The comparison table created using these parameters is given the section IV. 2.1 Methodology and Indexing The ontology techniques or methods used are one of the main criteria for effectively achieving semantic web. The methodologies include crawlers based on mining tools, inference engines, annotation tools etc. The accepted ontology description languages include RDF and OWL in their structure. These languages are necessary for the description of the ontologies. This comparison has technologies which include RDF, DBpedia, and SPARQL etc. Indexing is the process of storing the given data in an organised manner which helps in improving the speed of data retrieval operations. Without the use of indexing, a search engine will have to scan through all the documents in the database. Hence, the application of indexing to the content found during the crawling process eases up the retrieval task. Following are the types of indexing which can be found.