Similarity Search Techniques in Exploratory Search: A Review

Abstract

The past decade has seen a dramatic increase in the amount of data captured and made available to users for research. This increase amplifies the difficulties users' face in finding the data most relevant to their information needs. The document similarity search is one of the most important topics in the field of information science, especially due to the popularity of the internet applications that deal with unstructured data sources such as World Wide Web. Efficiency of similarity search has become one of the most important issues. A typical example of similarity search is in multimedia databases that manage objects without structure, i.e. images, fingerprints or audio clips. Here similarity search is involved in retrieving the most similar fingerprint to a given one. Another example is in text retrieval which is present in many systems, from simple text editors (finding words similar to a given one to correct edition errors) to big search engines (retrieving relevant documents for a given query). This study explores the use of similarity search for text data in the form of a brief review using the interface provided as a service after content-based searches has been performed. The findings will give us ideas as to how to incorporate similarity searches within others search engine architecture. © 2018 IEEE.