Classification of high dimensional Educational Data using Particle Swarm Classification

This paper explores the effectiveness of Particle Swarm Classification technique for tackling a classification problem in an emergent data mining field, called Educational Data Mining. More specifically, it applies Particle Swarm Classification to classify a data set of teachers´ classroom questions into the cognitive levels of Bloom´s taxonomy. Furthermore, the high dimensionality of questions data set enables investigating the effectiveness of particle swarm classification for the classification in high dimensional domains. In doing so, a data set of teachers´ classroom questions has been collected and annotated manually with Bloom´s taxonomy cognitive levels. Preprocessing steps have been applied to convert questions into a suitable representation. Using this data set, the effectiveness of Particle Swarm Classification has been evaluated and compared with four conventional machine learning techniques. The results show that Particle Swarm Classification is promising for tackling classification tasks in Educational Data Mining. Moreover, the results confirm that Particle Swarm Classification with proper confinement mechanism is effective for the classification in high dimensional domains. These conclusions are evidenced by the superior performance of Particle Swarm Classification over the four conventional machine learning techniques.