A neural network-based model for classifying the opinions of airline passengers published on social media

Analyzing customer opinions is one of the important approaches for business development. Considering the expansion and diversity of airlines, an accurate understanding of passengers’ opinions can be an effective step in surpassing companies in the competitive market. Therefore, a model for categorizing airline customers’ thoughts utilizing the Twitter platform was attempted to be presented in this study. This study's suggested model is predicated on two primary methods: Neural networks and text analysis. For this purpose, two layers of CNN and DNN were combined as the main layers in three different dimensions, and six models were formed. Then, by carrying out a case study and looking at several assessment indices, it was discovered that the models’ accuracy declined as their dimension rose. In other words, one-dimensional models are more accurate than two- and three-dimensional models in classifying textual data. On the other hand, among one-dimensional models, the CNN-based model has less error. Therefore, the one-dimensional CNN model was suggested as a proposed model to classify the opinions of airline passengers.