A Machine Learning–Based Sentiment Analysis Framework for Assessing and Predicting Teacher Performance Using QEC Data

Abstract

Data of any organization or institute is an important source for future planning and decision-making. Using Educational Data Mining we can explore educational data by applying different data mining and techniques of machine learning. It is a well-thought-out research topic that supplied fundamentals for important facts of learning and teaching procedures to improve education. The primary goal of every academic institution is to offer students a high-quality education. As a result, the primary goal of this research is to uncover information that will aid in improving the quality of education for children. High-level quality in education can be achieved by acquiring knowledge, which is obtained by the data related to students, teacher courses, and institutes. Many patterns can be discovered which help to make better decisions in the education sector. The experimental data consists of students' evaluation forms obtained by the QEC and filled by the students of an institute. We used the WEKA and PYTHON data mining tool for the analysis of the dataset. We used J48, Naïve base, RT (Random-Tree), and RF (Random-Forest) algorithms in WAKA for the evaluation of teacher performance and predict how to perform well. Then we used python as a tool and NLTK and Text Blob or Vader sentiment Library for the sentimental analysis of student's opinions. In our experiment, we achieve 95.56% accuracy and acceptable kappa statistics of 0.8038 using random forest. And the results showed that reviews have a prime role in the evaluation the educational data. This research provides the best direction for the development of the education field.