Text mining and sentiment analysis of teacher performance satisfaction in the virtual learning environment

Abstract

Although it is true that artificial intelligence and data science have become key tools that contribute to the improvement of many processes, identifying patterns and contributing to decision making, however, there are environments in which they are not yet being using it relevantly and effectively. The objective of this study is to identify the relevant factors, based on the opinions expressed by the students through the social network Twitter regarding the perception of satisfaction with the teaching performance during the virtual learning environment. For which sentiment analysis and text mining are used under the Python programming language environment, through JupyterLab. As results, it was determined that a predominance of 57.27% of positive polarity, identifying that the relevant factors of student satisfaction with teaching performance, are related to the development of the teacher in the class sessions that contributes to the learning of the process control subject through the use of simulation tools such as simulink and tools linked to proportional integral derivative (PID) controllers; on the other hand, there is a percentage of negative polarity of 15.45% that belongs to the factors linked to the laboratory sessions in which graphic representation and block diagrams were used to explain the class session.

Student satisfaction in the context of hybrid learning through sentiment analysis

“…Regarding the method used for the collection, storage, and processing of data, it is shown in Figure 1, which was validated through what was developed by Chamorro-Atalaya et al [38], in which he initially defines the way in which the data will be stored and collected through the social network Twitter, to then continue with the identification of feelings in which the levels of student satisfaction are described based on three categories, we define as "positive", "negative", and "neutral", all supported by the "nltk" and "vader" libraries contained in the Python programming software. Then a data preprocessing is carried out, with which the opinions generated by the students from Twitter can be cleaned, converting all the texts to lowercase, eliminating repeated words, and eliminating excessive spaces between words.…”