Classification of tweets for sentiment and trend analysis

Education and work-related reading have probably gone up and on-screen reading also increased considerably. A survey by the Pew Research Center in the year 2012 proved that the increase in ownership of tablet computers and electronic book reading devices, such as ipad, kindle fires, Nook, etc., the online book reading also increases considerably. A survey by National Literacy Trust during the year 2013 reported that more than 52% of students prefer to read electronic devices compared to 32% who preferred printed materials. The survey made by the Wall Street Journal, USA today and United Parcel Service (UPS) says that there is a rapid growth in online shopping during recent years. As online reading and online purchase have increased over the period, recommendation becomes mandatory for today's online world, an especially recommendation of books become most important for the student community, because of the availability of huge volume of books online. In order to gain the trusted recommendation people today depends on social networks. Twitter a social network is a popular microblogging service where users create tweets that sometimes express opinions about different topics which helps them to identify the current trends. The work proposed in this paper move towards recommending books based on trending topics from the trusted users obtained through provenance. The proposed work utilizes Naïve Bayesian technique and Moving Average (MA) method to identify the sentiments from tweets and trend respectively.