Issuer Credit Rating Performance Report Using Sentiment Analysis

Abstract

Indonesian Credit Rating Agency (CRA) is currently on its way to becoming the early mover of digital transformation. CRA controls macroeconomics and has a significant impact on many industries across the world. However, there are always those that can exploit it through asymmetric information and human interaction. A solution to reduce human interaction and enhancement is to build Natural Language Processing (NLP) sentiment analysis models and then display the results using an interactive dashboard story. Objectives are created for the aim of the project to be able to conduct a feasibility study, develop a model based on a press release dataset, conduct model evaluation, and display the results on an interactive dashboard. The research aims to utilise press release documents with NLP sentiment analysis to produce prescriptive analysis with interactive visualisation as the final output. Press release files are processed by using several Machine Learning (ML) algorithms such as Support Vector Machine (SVM), Multinomial Naive Bayes (MultinomialNB), Logistic Regression (LR), and Multi-Layer Perceptron Artificial Neural Network (MLP-Ann). This research will be carried out under Dynamic Systems Development (DSDM) and Knowledge Discovery Database (KDD). This will allow the researchers to achieve all objectives, permit models to perform very well, and let the output get displayed on a dashboard as a storyboard.