Dealing with Data from Multiple Web Sources
Natércia A. Batista, Michele A. Brandão, Michele Brito Pinheiro, Daniel Hasan Dalip, Mirella Moura Moro · 2018
Web data are heterogeneous and unstructured, which defines challenges for data crawling, integration and preprocessing. Different studies are "data-oriented" (i.e. based on the available data) but their results are restricted to their specific data. In contrast, there are various problems prior to identifying what data is needed to solve them, and often multiple data sources are needed. In this context, crawling, integrating and preprocessing data appropriately enables to create datasets for solving such problems. Therefore, this short course addresses these three activities by discussing challenges and practical solutions.