Textual Sentimental Classification using Convolution Neural Network Algorithm
K. Kavitha, Suneetha Chittineni, Naveen Kumar Andhavarpu, Vasundhara Chappa · 2022
Sentiment analysis on classification of text data is one of the emerging tasks in natural language processing compared to others. In the main, there is a need for big dig for meaningful information from the data present on the internet through sentiment analysis. Deep learning has already been successful in keeping up with sentiment analysis tasks utilizing deep learning models. Nowadays, client happiness is the most crucial factor for every mature firm. As a result, many businesses live and promote their products and services on social media platforms, where they eventually receive evaluations and feedback from their customers. Studying each word and phrase one at a time is time-consuming. Thus, evaluating the sentiment of all texts provides organizations with an overview of how positive and negative users are on a given issue. In this study, we investigated word embedding architecture utilizing the convolutional neural network algorithm or approach (CNN). We created a convolutional neural network architecture using a continuous loop of the tensor flow technique, along with word2vec and evaluated the outcome by using the measures accuracy, precision, recall, and f1-scoring. It reaches an accuracy of more than 92 percent, significantly boosting text categorization accuracy.