Dataset Summarization

Abstract. Recent research has shown the Linked Data cloud to be a potentially ideal basis for improving user experience when interacting with Web content across different applications and domains. Using the explicit knowledge of datasets, however, is neither sufficient nor straightforward. Dataset knowledge is often not uniformly organized, thus it is generally unknown how to query for it. To deal with these issues, we propose a dataset analysis approach based on knowledge patterns, and show how the recognition of ...

We present a Linked Data analysis method which relies on knowledge patterns for constructing a logical architecture of the knowledge in a dataset. This can then be exploited to compare heterogeneous datasets, enhance interoperability between them and make implicit knowledge emerge.

We introduce a family of unsupervised, domain-free, and asymptotically optimal model-independent algorithms based on the principles of algorithmic probability and information theory designed to minimize the loss of algorithmic information, and thereby avoiding certain deceiving phenomena and distortions known to occur in statistics and entropy-based approaches. Our methods include a lossless-compression-based lossy compression algorithm that can select and coarse-grain data in an algorithmic-complexity fashion (without the use of popular compression algorithms) by collapsing regions that may procedurally be regenerated from a computable candidate model. We show that the method can perform dimension reduction, denoising, feature selection, and network sparsification, while preserving the properties of the objects. As validation case, we demonstrate the methods on image segmentation against popular methods like PCA and random selection, and also demonstrate that the method preserves the graph-theoretic indices measured on a well-known set of synthetic and real-world networks of very different nature, ranging from degree distribution and clustering coefficient to edge betweenness and degree and eigenvector centralities, achieving equal or significantly better results than other data reduction and the leading network sparsification methods (Spectral, Transitive).

Recent research has shown the Linked Data cloud to be a potentially ideal basis for improving user experience when interacting with Web content across different applications and domains. Using the explicit knowledge of datasets, however, is neither sufficient nor straightforward. Dataset knowledge is often not uniformly organized, thus it is generally unknown how to query for it. To deal with these issues, we propose a dataset analysis approach based on knowledge patterns, and show how the recognition of patterns can support querying datasets even if their vocabularies are previously unknown. Finally, we discuss results from experimenting on three multimedia-related datasets.

Recent research has shown the Linked Data cloud to be a potentially ideal basis for improving user experience when interacting with Web content across different applications and domains. Using the explicit knowledge of datasets, however, is neither sufficient nor straightforward. Dataset knowledge is often not uniformly organized, thus it is generally unknown how to query for it. To deal with these issues, we propose a dataset analysis approach based on knowledge patterns, and show how the recognition of patterns can support querying datasets even if their vocabularies are previously unknown. Finally, we discuss results from experimenting on three multimedia-related datasets.

Recent research has shown the Linked Data cloud to be a potentially ideal basis for improving user experience when interacting with Web content across different applications and domains. Using the explicit knowledge of datasets, however, is neither sufficient nor straightforward. Dataset knowledge is often not uniformly organized, thus it is generally unknown how to query for it. To deal with these issues, we propose a dataset analysis approach based on knowledge patterns, and show how the recognition of patterns can support querying datasets even if their vocabularies are previously unknown. Finally, we discuss results from experimenting on three multimedia-related datasets.

A distributed, hierarchical information service for computer networks might use several service instances, located in different layers. A distributed directory service, for example, might be comprised of upper level listings, and local directories. The upper level listings contain a compact version of the local directories. Clients desiring to access the information contained in local directories might first access the high-level listings, in order to locate the appropriate local instance. One of the keys for the competent operation of such service is the ability of properly summarizing the information which will be maintained in the upper level directories. We analyze the case of the Lookup Service in the Information Services plane of perfSONAR performance monitoring distributed architecture, which implements IP address summarization in its functions. We propose an empirical method, or heuristic, to perform the summarizations, based on the PATRICIA tree. We further apply the heuristic on a simulated distributed test bed and examine the results.

Recent research has shown the Linked Data cloud to be a potentially ideal basis for improving user experience when interacting with Web content across different applications and domains. Using the explicit knowledge of datasets, however, is neither sufficient nor straightforward. Dataset knowledge is often not uniformly organized, thus it is generally unknown how to query for it. To deal with these issues, we propose a dataset analysis approach based on knowledge patterns, and show how the recognition of patterns can support querying datasets even if their vocabularies are previously unknown. Finally, we discuss results from experimenting on three multimedia-related datasets.