Automatic Theme Extraction for Marine Data Marts Using Semantic Web

Provided by: AICIT
Topic: Big Data
Format: PDF
Due to multi-source, polymorphism and diversity of marine data, making use of existing technology to establish the marine data mart has become a serious problem. In addition, how to automatically and fast create data mart with few users' requirements and little background knowledge about marine has been paid much attention. In this paper, the authors propose a novel method to extract theme for marine data marts based on semantic web. They obtain the similarities between field names in the marine data warehouses; then, utilize the spectral clustering algorithm to partition field names and obtain field clusters to extract themes.

Find By Topic