Classification of Texts with Neural Networks

Author: Gela Besiashvili
Co-authors: P.Qarchava
Keywords: convolutional and recurrent neural networks, sentiment Analysis
Annotation:

Recurrent neural networks were applied for processing the texts (analyze, categorize, classification) mostly, in contrast with the convolutional neural networks that were primarily used for classifying the multimedia information. However, research showed that convolutions can successfully be applied in classification of texts as well. Convolutions enable us to consider words’ context in learning process, which solves the main problem of classifying the text – loss of context and improves the result at the same time.

Lecture files:

Classification of Texts with Neural Networks [en]
ტექსტების კლასიფიკაცია ნეირონული ქსელებით [ka]