Inhalt des Dokuments
Bachelor Thesis: A hybrid approach for emotion-based sentiment analysis for Twitter data
A hybrid approach for emotion-based sentiment analysis for Twitter data
In this thesis we use a hybrid approach for a sentiment analysis for Twitter documents. This approach uses a lexical approach to label the data based on emotional keywords and emojis. After that the data is cleaned with multiple preprocessing techniques and then converted into vectors with the help of Doc2Vec. In the end this data is used to create three different machine learning models, based on linear regression, svm and knn, that are able to classify tweets into one out of six different basic emotions: anger, disgust, fear, happiness, sadness and surprise.
The machine learning algorithms will be compared with each other based on their F1 scores. We compare how the classifiers perform overall and how good they perform for a certain class.
Supervisor: Katerina Katsarou 
Type: Bachelor Thesis
Duration: 4 months
10587 Berlin, Germany
Phone: +49 30 8353 58811
Fax: +49 30 8353 58409
e-mail query