Making Teaching and Learning Effective Using Analytics

Abstract

In this paper, an empirical comparison of three state-of-the-art classifier methods (artificial immune recognition systems, Lazy-K Star, and random tree) to predict teachers’ ability to adapt in a classroom environment is carried out. Two educational databases are used for this task. First, measures collected in an academic context, especially from classroom visits, are used (database 1). Then, the three classifiers quantify the acts, behaviors, and characteristics of teaching effectiveness and the teacher’s “ability to adapt in the classrooms.” Professional classrooms visits to more than 200 teachers are used as the second database (database 2). An interactive grid gathering 63 educational acts and behaviors is conceived as an observation instrument for those visits. Within the Waikato Environment for Knowledge Analysis library environment, and with the progressive enhancement of the raw database, the utilization of state-of-the-art classification methods when predicting teaching effectiveness shows promising results, especially when data quality issues are considered.