In order to illustrate the method by which online education services are provided, this article makes use of service blueprint technology. The service process inside the online learning system, the customer touch point, and the consumer-facing features of the service are all shown here. When describing services, using this top-down approach helps avoid problems associated with transience. We proposed sentiment analysis on educational tweets using NLP with WORDNET. The dataset has normalized using Part of speech (POS) tagging with Natural Language Processing (NLP) and WORDNET. Additionally, the method of Principal Component Analysis was utilized for Feature Reduction. Elastic net, recursive feature, and hybrid machine learning (decision tree (DT) and random forest (RF)) were utilized during Ensemble Feature Selection. Finally, Linear Support Vector Machine (SVM), Gaussian naive bayes (GNB), and Linear Regression (LR) were utilized as ML classification algorithm. The experimental results have shown with different performance metrics.

Ensemble Feature selection, POS, Sentiment Analysis, Education, NLP, WORDNET