Article Preview
Top1. Introduction
In today’s modern era where the state of media is changing constantly, social media has gained immense popularity and social media revolution has come into play (Gupta and Nitin, 2017). Social media involves blogs, forums and various other aspects of interactive presence that enable the individuals to engage in conversations or discussions over a particular news article, blog post or event. As a result, social media provides a way to increase reachability of ideas, views or content of any other form. Lots of research works have been executed to solve problems in the domain of online social media (Dey, Borah, Babo, & Ashour, 2018). This includes the studies to compare hashtags used in Instagram and Twitter (Highfield and Leaver, 2015), examine the demographics of social media users (Davenport et al., 2014), determine challenges when accessing the Twitter data (Kelley et al., 2013), differentiate users on how they react on Instagram and examine the full photo content (Hu et al., 2014; Mittal et al., 2017), understand people’s behavior through their speech of text on social media (Schwartz and Ungar, 2015), analyze the comments made by the user and his friends (Ko et al., 2014), to compare various tools and techniques for social media analytics (Batrinca and Treleaven, 2015) and recommend keywords to the users (Beliga et al., 2015).
A problem that has not been much discussed yet is regarding the selection of a social media platform most appropriate for a specific content to be posted. Each social media content need not be posted on all social media sites and usually, a content uploader is unable to decide the following while posting content- Which social media platform is most desirable to post different content? What content should be posted, i.e., which terms or keywords can boost the online spread of a post or increase the post engagements? Therefore, this paper proposes an approach to determine the most suitable social media platform (Facebook/ Twitter) based on post content and also the most likable content (or key terms) to ensure a wide and quick spread of a post, i.e. to make the post viral.
According to the studied literature, few research gaps exist in relevance to the objective of this paper, i.e., in the direction to enhance the influence of a post. There has been a statistical analysis in some papers related to the number of likes, comments, shares, retweet, status count and others (Davenport et al., 2014; Schwartz and Ungar, 2015; Ko et al., 2014; Kelley et al., 2013). Whereas, the other papers have reported their findings based on the surveys conducted on a sample of users. These gaps require an effective approach to identify the most relevant social media platform and keyword suggestion approach for influence maximization. A comparative analysis of various social media applications is presented in the existing literature (Ko et al., 2014). Davenport et al. worked on demographic details to find out the most preferable social media platform (Davenport et al., 2014). Therefore, based on this studied literature, we identified a few research gaps which motivated to work further in this direction. No previous work exists to identify of most preferable social media platform based on historical posts content of similar theme and users’ reaction on these posts.
This research work is done to get the answer to the following research questions:
RQ1: Is it possible to measure the most preferable social media application for a specific theme/ topic? Or which social media application will make the topic more popular as compared to other social media?
RQ2: Is it possible to measure theme influence based on users’ engagements and keywords used in the discussion of the post?
In previous works, keyword/ theme-based influence identification has been performed based on hashtags (Highfield and Leaver, 2015), single term frequency (Beliga et al., 2015) and n-gram word extraction, etc. However, a single approach solely will not be able to handle this issue.