Using Signal processing to Investigate User Web Search Behaviour on Topics of Interest with Multiple Periodicities
Jivashi Nagar, Hussein Suleman · 2019
A Web user searches for multiple topics of interest on the Web on a regular basis. The topics may be searched periodically at a particular time or at different times, creating temporal patterns with different periodicities in the search history. To improve a user’s Web search experience, the multiple periodicities of the topics of interest of a user can be exploited. This study proposes to find multiple periodicities of a user’s topic of interest through signal processing. It is found that Fast Fourier Transform can be used to find multiple periodicities of a topic as well as to predict the temporal pattern of the topics with accuracy and error depending on the training data size.