Capturing Changes and Variations from Teachers’ Time Series Usage Data

Abstract

The type of time series data is very common in educational settings, such data set usually provide usage trajectories involved complex contexts and unknown noisy. In this paper, we describe a statistical analysis method to detect a switch point over teacher’s comment data set which is collected from a student evaluation system. We find that it is indeed possible to employ a statistical model and Bayesian inference to detect teacher who increased, decreased or no-changed their behavior. We also illustrated the potential of the application of change detection to conduct data exploration, which can be relative to understand teacher technology adoption and usage transition.