Interactive hierarchical tag clouds for summarizing spatiotemporal social contents

Abstract

Abstract-In recent years, much effort has been invested in analyzing social network data. However, it remains a great challenge to support interactive exploration of such huge amounts of data. In this paper, we propose Vesta, a system that enables visual exploration of social network data via tag clouds. Under Vesta, users can interactively explore and extract summaries of social network contents published in a certain spatial region during a certain period of time. These summaries are represented using a novel concept called hierarchical tag clouds, which allows users to zoom in/out to explore more specific/general tag summaries. In Vesta, the spatiotemporal data is split into partitions. A novel biclustering approach is applied for each partition to extract summaries, which are then used to construct a hierarchical latent Dirichlet allocation model to generate a topic hierarchy. At runtime, the topic hierarchies in the relevant partitions of the user-specified region are merged in a probabilistic manner to form tag hierarchies, which are used to construct interactive hierarchical tag clouds for visualization. The result of an extensive experimental study verifies the efficiency and effectiveness of Vesta.

Automatic transfer learning for short text mining

“…Specifically, when the training data in the target task are insufficient for a good data modeling, it transfers the useful knowledge from the related auxiliary data which are from another task to enrich the data features. In this case, more data characteristics are integrated into the data learning facilitating an improved learning results [1][2][3][4][5][6].…”