Page 322 - Data Science Algorithms in a Week

Predictive Analytics for Thermal Coal Prices Using Neural Networks … 303

supply, and these in turn are determined by both quantifiable and non-quantifiable
variables that are the model of their price actually.
The thermal coal market has another characteristic: despite being an oligopoly, it is
not possible to be strategic in terms of prices, as is the case with oil. Coal companies
almost always have to be reactive with respect to market events. This is a phenomenon
that Peter Senge (2006) describes: The companies that are reactive in the markets begin
to worry about the facts, and concern for the facts dominates business deliberations such
as the price of coal of last month.
To analyze the coal market, we formed a panel of experts from around the world. The
Delphi methodology was used to investigate with this panel of experts which are the most
important strategic variables globally that influence the price of thermal coal. Once a
consensus is reached, AI techniques can be used to verify these variables and build
predictive models to calculate the price of thermal coal. This prediction can provide
strategic support to the coal mining companies (Phil, 1971).
In the history of the prices of thermal coal, there exists the following milestones that
have marked great fluctuations (Ellerman, 1995; Yeh & Rubin, 2007; Ming & Xuhua,
2007; Finley, 2013; EIA, 2013):

 Oil crisis of the 1970s - This crisis caused countries to rethink their dependence
on oil for power generation and gave impetus to coal as an alternative;
 Emphasis on sea transportation - Developers of mining projects dedicated
mainly to exports, promoted the development of the market for coal transported
by sea and therefore globalized the supply and demand of coal (previously the
coal was consumed near places where it was extracted)
 Prices indices for Coal – The creation of price indices at different delivery
points (FOB Richards Bay, CIF Rotterdam) that gave more transparency in
transactions and helped better manage market risk;
 Industrialization of emerging economies (especially China) – This
industrialization gave support to demand never seen before
 Emergence of financial derivative markets – This financial markets offered
more tools to manage price risk (they also promoted the entry of new players,
such as banks)
 Global warming and climate change - The publication of studies on global
warming and climate change that led countries worldwide to take action to
reduce CO2 emissions and thus reduce the use of coal
 Germany shuts down all its nuclear plants – This happened after the accident
at the Fukushima Nuclear Plant in Japan in March 2011, indirectly driving an
increase in power generation with renewable, Coal and Natural Gas