Department of Computer Science
Visualizing global manifold based on distributed local data abstractions
Mining distributed data for global knowledge is getting more attention recently. The problem is especially challenging when data sharing is prohibited due to local constraints like limited bandwidth and data privacy. In this paper, we investigate how to derive the embedded manifold (as a 2-D map) for a horizontally partitioned data set, where data cannot be shared among the partitions directly. We propose a model-based approach which computes hierarchical local data abstractions, aggregates the abstractions, and finally learns a global generative model - generative topographic mapping (GTM) based on the aggregated data abstraction. We applied the proposed method to two benchmarking data sets and demonstrated that the accuracy of the derived manifold can effectively be controlled by adjusting the data granularity level of the adopted local abstraction.
Data visualization, Data privacy, Data mining, Bandwidth, Computer science, Aggregates, Automatic control, Memory, Parametric statistics, Covariance matrix
Source Publication Title
Proceedings of the Fifth IEEE International Conference on Data Mining (ICDM’05)
Han, Jiawei ; Wah, Benjamin W. ; Raghavan, Vijay ; Wu, Xindong ; Rastogi, Rajeev
Houston, United States
Copyright © 2005 by The Institute of Electrical and Electronics Engineers, Inc.
This work is jointly supported by RGC Central Allocation Research Grant (HKBU 2/03/C) and Hong Kong Baptist University FRG Grant (FRG/05-06/I-16).
Link to Publisher's Edition
Zhang, X., & Cheung, W. (2005). Visualizing global manifold based on distributed local data abstractions. Proceedings of the Fifth IEEE International Conference on Data Mining (ICDM’05). https://doi.org/10.1109/ICDM.2005.150