Final project for the course Cognitive Computational Modeling of Language and Web Interaction, offered at Radboud University, Nijmegen.
The project is titled "String Kernels in Sentiment Analysis". The goal is to compare feature-light string kernel methods to more commonly used bag-of-word-models and word embeddings using linear or RBF kernels with SVM. Check out full report for more details.