Decoding User Feedback: A Convolutional Neural Network Approach to Sentiment Analysis in Event Management
Muhammad Riefky Hadid, Casi Setianingsih, Roswan Latuconsina, Ari Cahya Saputra, Ariana Novanti, Muhammad Izzudin Islam · 2023
In a committee event, the organizer usually evaluating their event once the event is done. One way to evaluate an event is to ask the participant to write a feedback about the event. The amount of feedback that comes in will impact on the time it takes to analyse it. The more feedback comes in, then it will take more time for the committee to analyse it. With Machine Learning and Deep Learning, feedbacks that comes in will be analyse with short amount of time and can be determined the sentiment value. In this research, a sentiment analysis model will be designed using Convolutional Neural Network (CNN) classification method. The output of this sentiment analysis model will be determining whether the sentiment value of a feedback is positive or negative. This sentiment analysis model later will be tested with Performance Evaluation. There are 4 parameters that will be tested, Accuracy, Recall, Precision, and F1 Score. The result of this research hopefully will be helpful to determine the overall feedback that comes in on an event with short amount of time. From the result of the test, the model with the highest accuracy of 94,4% is model with data partition of 70:30, epoch value of 100, batch size value of 32, and learning rate value of 0,00003.