Co-Clustering by Bipartite Spectral Graph Partitioning for Out-of-Tutor Prediction.
Shubhendu Trivedi, Zachary A. Pardos, Gábor N. Sárközy, Neil Thomas Heffernan · 2012
Learning a more distributed representation of the input fea-ture space is a powerful method to boost the performance of a given predictor. Often this is accomplished by parti-tioning the data into homogeneous groups by clustering so that separate models could be trained on each cluster. In-tuitively each such predictor is a better representative of the members of the given cluster than a predictor trained on the entire data-set. Previous work has used this basic premise to construct a simple yet strong bagging strategy. However, such models have one significant drawback: In-stances (such as students) are clustered while features (tu-tor usage features/items) are left alone. One-way cluster-ing by using some objective function measures the degree of homogeneity between data instances. Often it is noticed that features also influence final prediction in homogeneous groups. This indicates a duality in the relationship between clusters of instances and clusters of features. Co-Clustering simultaneously measures the degree of homogeneity in both data instances and features, thus also achieving clustering and dimensionality reduction simultaneously. Students and features could be modelled as a bipartite graph and a si-multaneous clustering could be posed as a bipartite graph partitioning problem. In this paper we integrate an effective bagging strategy with Co-Clustering and present results for prediction of out-of-tutor performance of students. We re-port that such a strategy is very useful and intuitive, even improving upon performance achieved by previous work.