Recommending Environmental Big Data Using Semantically Guided Machine Learning - Large Scale and Big Data: Processing and Management - page 465

15.1.3 l inkeD o Pen D ata a PProaCh
Next-generation recommendation systems on the web would be based around the
Linked Open Data (LOD). This chapter deals with five big environmental data
sources to comprehend and discover knowledge in an integrated fashion, interprets
knowledge, and finally proposes a semantically guided machine learning-based
architecture to recommend knowledge from that discovery. To make the knowledge
widely accessible it is important to publish the interpreted recommendation on the
LOD. Publishing recommendations about the Big Data integrated on the LOD will
make analytical knowledge about the data machine readable and autonomously
available to the machines. Prior analytical recommendations about the Big Data
would provide an automatic framework that could be used to prioritize, optimize,
and minimize Big Data accessibility issues. It would be interesting to capture some
of the basic understanding about the LOD concept before moving toward the detailed
sections about the Big Data recommendation system. Primarily web pages are built
in hypertext markup languages (HTML) that make the web as a web of document.
All the webs of documents are possible to navigate through the hyperlinks. In recent
years, the World Wide Web is moving from the web of document to the web of data
[3,4,6]. In this situation, metadata is moving from merely machine-readable toward
automatic processing power, where it is essential to break the record and repositories
silos to make data (especially metadata) into machine understandable pieces. These
pieces could be presented using Resource Description Framework (RDF), which
provides a data model for presenting metadata in a form that can be understood
and process into triple statements (i.e., subject, object, predicate) automatically, by a
machine [30-32,40]. LOD relies on documents containing data in the RDF format.
However, rather than simply connecting these documents, linked data uses RDF to
make typed statement that links arbitrary things in the world, which may be referred
to as web of things. Tim Berners-Lee designed a set of rules for publishing data
on the web in a way that all published data become part of a big machine readable
umbrella. The rules were

• Use Universal Resource identifier (URI) as names of things

• Use HTTP URIs so that people can look up those names

• When someone looks up an URI, provides useful information, using the

standards (RDF, SPARQL)

• Include links to the other URIs, so that they can discover more things

They have become known as the LOD principles and provide the basic idea for
publishing and connecting data on the web. Although the idea of LOD has yet to be
accepted as mainstream (like the web WWW is known to all), there are a lots of
LODs already available. The so called LOD cloud covers more than an estimated 50
billion facts from many different domain like geography, media, biology, chemistry,
economy, energy, etc. [3,4,6,31,32]. The data from the cloud can also be reused for
building the end users applications or for the research purposes [11]. The basic idea
of a semantic web is to provide cost-effective ways to publish information in a dis-
tributed environment (Figure 15.1). The LOD fulfills these ideas where organizations