Application of the Convolution Neural Network in the Text Sentiment Analysis
Application of the Convolution Neural Network in the Text Sentiment Analysis

Author(s): Jasmina Novaković, Suzana R. Marković
Subject(s): Social Sciences
Published by: Udruženje ekonomista i menadžera Balkana
Keywords: Sentiment analysis; CNN; Neural networks; Machine learning; Classification
Summary/Abstract: Sentiment analysis deals with the analysis of the opinions, attitudes and emotions of the people who wrote a certain text. The goal of text sentiment analysis is to detect positive, neutral or negative sentiments from the text. The ability to automatically recognize emotions from text has many practical applications, such as customer sentiment analysis, social media monitoring, and customer feedback analysis. In this paper, we considered the application of neural networks in text sentiment analysis, using the Python programming language and Keras as a deep learning Python library. In the experimental part of the work, we trained a convolution neural network to classify text into different classes depending on the recognized emotions. We obtained a model that we tested on a test data set and considered the classification accuracy of the obtained model.

• Details
• Contents