Should I Care about Your Opinion? Detection of Opinion Interestingness and Dynamics in Social Media

Abstract

In this paper, we describe a set of reusable text processing components for extracting opinionated information from social media, rating it for interestingness, and for detecting opinion events. We have developed applications in GATE to extract named entities, terms and events and to detect opinions about them, which are then used as the starting point for opinion event detection. The opinions are then aggregated over larger sections of text, to give some overall sentiment about topics and documents, and also some degree of information about interestingness based on opinion diversity. We go beyond traditional opinion mining techniques in a number of ways: by focusing on specific opinion-target extraction related to key terms and events, by examining and dealing with a number of specific linguistic phenomena, by analysing and visualising opinion dynamics over time, and by aggregating the opinions in different ways for a more flexible view of the information contained in the documents.

Ontology for cultural variations in interpersonal communication: Building on theoretical models and crowdsourced knowledge

“…GATE is the biggest open‐source project for textual annotation that allows the use of custom ontologies (Cunningham et al, ). GATE is also used in all types of human computational tasks ranging from detecting events from financial text (Hogenboom, Hogenboom, Frasincar, Schouten, & van der Meer, ) to detecting opinions from social media (Maynard, Gossen, Funk, & Fisichella, ). It has one of the largest and most diverse user community for any text engineering tools (Cunningham et al, ).…”

A framework for real-time semantic social media analysis

“…via TwitIE [16], as well as sentiment analysis (detecting whether a social media post is opinionated, what kind of opinion is expressed, who the holder of the opinion is, what the opinion is about, and so on) [17,18]. Where appropriate, entities and terms are associated with relevant URIs from Linked Open Data (LOD) via YODIE [19].…”

“…Figure 6 shows a choropleth depicting the distribution of MPs' tweets which discuss the UK economy (the most frequent theme) during the week beginning 2 March 2015. This is a dynamic visualisation, based on the Leaflet library 17 and the aggregated query results returned by Mímir for each theme and NUTS1 region. The choropleth has a pull-down menu from which the user can select the topic of interest, and this re-draws the map accordingly.…”

SentiFlow: An Information Diffusion Process Discovery Based on Topic and Sentiment from Online Social Networks

“…There exist some studies in information diffusion based on sentiment [14][15][16], but they do not consider the opinion flow between communities. Other researchers analyzed how to visualize topics and opinions in SNS [17][18][19][20][21], but they lack in the information process flow. In this research, a new semantic hidden Markov model (HMM) for discovering information diffusion, named SentiFlow, is introduced to discover probabilistic information flow in consideration of topics and sentiment.…”