Opinion mining analytics of IoT ecosystem by Profile of Mood State with Logistic Regression

Abstract

Internet of Things continues to redefine modus operandi across diverse socioeconomic and professional domains thereby generating an un-abating global discuss on the adoption and functionalities of smart devices. Since emotions play a critical role in decision making according to the psychological domain of emotion science, the paramount importance of periodic delineation of stakeholders' mood is imperative for policy makers. Whereas opinion mining analytics of IoT discussions have returned state-of-the-arts, there is need to address germane factors seldom factored into existing literatures. This study therefore consolidates on current frameworks through a bi-modal descriptive and content-based analytics of IoT ecosystem for detecting key mood domain and the BlueCheckCredibility status of IoT tweeters using Profile of Mood State and Nomogram-based analytics. With a 99.5% precision rate by Logistic regression model, social characteristic attributes of acquired ethnographic data points turns mutually exclusive to the credibility status of IoT opinion molders while tweet properties contributes higher discriminative tendencies for identifying negative IoT emotions. The impact of Internet of Things on data science is likewise unraveled through bi-gram content analytics to identify topical discussions encapsulated in the acquired tweet corpus.