Fundamentals of a graph transformation based web data processing system
Gabor Imre, Gergely Mezei · 2014
Internet is the most complex and complete source of information in the history of mankind. The innumerable webpages and the myriads of data providers form a complex, highly heterogeneous, continuously evolving system. This environment demands continuous research of information retrieval. Here we present our contribution: a lightweight, semi-formal approach of web exploration and web data analysis. Our approach focuses on analyzing heterogeneous, semi- or barely structured web data. Complex queries can be performed over multiple heterogeneous data sources; graph transformations can be used to adjust the queries and to analyze the results. This approach integrates the a priori human knowledge, an effective web querying method, the formality of graph transformations and optionally qualitative and quantitative analysis algorithms. We present an illustrative case study performing multi-domain search and data transformations: Stack Overflow users are searched with high contribution in a topic, then the LinkedIn profiles of these users is tried to be matched.