A Relevancy Classification System for Web Search Results in Mobile Devices

Mobile devices have become an increasingly popular means for the delivery of streaming audio, streaming video and internet content. However, users still face many limitations when searching for the most relevant, most popular, and often most useful video/audio source. This work presents a novel algorithm implemented in the form of an iPhone Application (i.e., App) useful in detecting and classifying the most relevantly searched You-Tube videos A video searching algorithm is first implemented using Google´s sophisticated You-Tube Application Interface (API) and is thus integrated with the Objective-C iPhone App. The You-Tube API provides a convenient mechanism for identifying a list of likely video candidates based on a set of search criteria as provided by the user. Next, viewer´s comments are extracted from each candidate and subsequently matched with a list of pre-specified key words. A score is computed from the matching algorithm resulting in a classification of the you-tube video. The algorithm is tested on the music video genre available at you-tube. Classification results are applied to 188 music videos as performed by The Beatles.