Sie befinden Sich nicht im Netzwerk der Universität Paderborn. Der Zugriff auf elektronische Ressourcen ist gegebenenfalls nur via VPN oder Shibboleth (DFN-AAI) möglich. mehr Informationen...
16th International Conference on Information Technology-New Generations (ITNG 2019), p.557-562
Ort / Verlag
Cham: Springer International Publishing
Link zum Volltext
Quelle
Alma/SFX Local Collection
Beschreibungen/Notizen
Sentiment analysis is a series of methods, techniques, and tools about detecting and extracting subjective information, such as opinion and attitudes, from language. The goal of our project was to classify movies’ reviews, by analyzing the polarity (positive or negative) of each paragraph in a review (Cui et al., Neurocomputing 187:126–132, 2016). We experimented with various RNN models on the Nervana Neon deep learning framework, an open-source framework developed by Nervana Systems, in order to improve accuracy in training and validation data. We experimented with network architecture, hyper parameters (batch size, number of epochs, learning rate, batch normalization, depth, vocabulary size) in order to find out which model works best for sentiment classification. This paper presents our findings and conclusions.