00000nam a22000004a 4500 UP-1686042739784911453 Buklod 20240520133554.0 m |o d | ta 240520s2024 xxu r |||| u|eng d (Buklod)UP-1685954869149757829 DSTC rda DSTC eng LG 995 2023 S8 L67 Lopez, Ma. Veronica V. author From Micro to Mainstream: Predicting the Mentions of Top Local Brands by Mainstream Media Twitter Accounts Ma. Veronica V. Lopez. Quezon City University of the Philippines February 2024 98 pages with colored illustrations ABSTRACT With the ubiquity of social media use among the general public, Brands need to pay more attention to how they are perceived online. One of the stronger influencers online are the same mainstream media players, the key figures of TV, print, and radio, who have likewise established strong followings on social media platforms, particularly Twitter/X. This study aims to predict whether a Brand will be mentioned by a Media Twitter account for a particular day, with elements of classical media theory as the basis for setting up predictor variables. Latent Semantic Analysis was used to determine the nature of the Media tweets, after which classification techniques Logistic Regression and Superior Vector Machines were used to predict the actual occurrence of Brand mention. Of the five Brands that were subjected to this analysis, a passable prediction model was generated for three Brands using Support Vector Machines. The variables based on traditional media theory also were not significant to the larger scope but one Brand showed indications where an extension of social media theory was significant to their Brand being mentioned. Classification. Latent Semantic Analysis. Logistic Regression. Media. Social Media. Support Vector Machines. Text Analysis. Twitter. FI Thesis. UPD DSTC LG 995 2024 S8 L67 Book