Sentiment Analysis of Lithuanian Texts Using Deep Learning Methods

We describe experiments in sentiment analysis of the Lithuanian texts using the deep learning methods: Long Short-Term Memory (LSTM) and Convolutional Neural Network (CNN). Methods used with pre-trained Lithuanian neural word embeddings are tested with different pre-processing techniques: emoticons restoration, stop words removal, diacritics restoration/elimination. Despite the selected pre-processing technique, CNN was always outperformed by LSTM. Better results (reaching an accuracy of 0.612) were achieved with the undiacritized texts and undiacritized word embeddings. However, these results are still worse if compared to the ones obtained using Support Vector Machines or Naive Bayes Multinomial and with the frequencies of words as features.