Sentiment strength detection for the social web

TLDR

Sentiment analysis traditionally targets commercial tasks, but its application to the social web—especially Twitter—faces challenges because existing algorithms rely on indirect cues that can be confounded by genre or topic, leading to spurious sentiment detection. This study evaluates an improved version of SentiStrength for detecting sentiment strength on the social web using direct sentiment indicators. The authors test SentiStrength 2, an enhanced algorithm that emphasizes direct sentiment cues rather than indirect indicators, across multiple social media platforms. Across six diverse social‑web datasets, SentiStrength 2 outperforms a baseline in both supervised and unsupervised settings, though it is sometimes outperformed by machine‑learning methods and shows weaker performance for positive sentiment in news‑related discussions, yet remains robust enough for broad application.

Abstract

Abstract Sentiment analysis is concerned with the automatic extraction of sentiment‐related information from text. Although most sentiment analysis addresses commercial tasks, such as extracting opinions from product reviews, there is increasing interest in the affective dimension of the social web, and Twitter in particular. Most sentiment analysis algorithms are not ideally suited to this task because they exploit indirect indicators of sentiment that can reflect genre or topic instead. Hence, such algorithms used to process social web texts can identify spurious sentiment patterns caused by topics rather than affective phenomena. This article assesses an improved version of the algorithm SentiStrength for sentiment strength detection across the social web that primarily uses direct indications of sentiment. The results from six diverse social web data sets (MySpace, Twitter, YouTube, Digg, Runners World, BBC Forums) indicate that SentiStrength 2 is successful in the sense of performing better than a baseline approach for all data sets in both supervised and unsupervised cases. SentiStrength is not always better than machine‐learning approaches that exploit indirect indicators of sentiment, however, and is particularly weaker for positive sentiment in news‐related discussions. Overall, the results suggest that, even unsupervised, SentiStrength is robust enough to be applied to a wide variety of different social web contexts.