Sentiment Analysis of German Twitter

TL;DR

This thesis advances German Twitter sentiment analysis by creating a new annotated corpus and developing novel, outperforming methods across lexicon generation, fine-grained opinion mining, message polarity classification, and discourse-aware analysis.

Contribution

It introduces a new German Twitter sentiment corpus and proposes innovative methods that surpass existing techniques in multiple sentiment analysis tasks.

Findings

Dictionary-based lexicons outperform corpus- and embedding-based ones.

The linear projection algorithm improves lexicon quality.

CRF models outperform RNNs in source and target prediction.

Abstract

This thesis explores the ways by how people express their opinions on German Twitter, examines current approaches to automatic mining of these feelings, and proposes novel methods, which outperform state-of-the-art techniques. For this purpose, I introduce a new corpus of German tweets that have been manually annotated with sentiments, their targets and holders, as well as polar terms and their contextual modifiers. Using these data, I explore four major areas of sentiment research: (i) generation of sentiment lexicons, (ii) fine-grained opinion mining, (iii) message-level polarity classification, and (iv) discourse-aware sentiment analysis. In the first task, I compare three popular groups of lexicon generation methods: dictionary-, corpus-, and word-embedding-based ones, finding that dictionary-based systems generally yield better lexicons than the last two groups. Apart from this, I propose a linear projection algorithm, whose results surpass many existing automatic lexicons. Afterwords, in the second task, I examine two common approaches to automatic prediction of sentiments, sources, and targets: conditional random fields and recurrent neural networks, obtaining higher scores with the former model and improving these results even further by redefining the structure of CRF graphs. When dealing with message-level polarity classification, I juxtapose three major sentiment paradigms: lexicon-, machine-learning-, and deep-learning-based systems, and try to unite the first and last of these groups by introducing a bidirectional neural network with lexicon-based attention. Finally, in order to make the new classifier aware of discourse structure, I let it separately analyze the elementary discourse units of each microblog and infer the overall polarity of a message from the scores of its EDUs with the help of two new approaches: latent-marginalized CRFs and Recursive Dirichlet Process.