4 Chapter 1
anomaly is associated, genetically and/or spatially, with mineral deposits of the type
sought, intersecting or integrated anomalies of various types are of interest in target
generation. The process of analysing and integrating such pieces of spatial geo-
information is called predictive modeling. This volume is further concerned with
predictive modeling of only geochemical anomalies and prospective areas.
This chapter explains the concepts of (a) predictive modeling, (b) predictive
modeling of geochemical anomalies and prospective areas and (c) application of a
geographic information system (GIS) in predictive modeling of geochemical anomalies
and prospective areas. A GIS consists of computer hardware, computer software,
geographically-referenced or spatial data sets and personnel.
WHAT IS PREDICTIVE MODELING?
To understand the concepts of predictive modeling of geochemical anomalies and
prospective areas via applications of GIS, it is imperative to define and understand what
model means. The Wiktionary (Wikimedia Foundation, 2007) defines model as “a
simplified representation (usually mathematical) used to explain the workings of a real
world system or event”. The Oxford English Dictionary (Oxford University Press, 2007)
defines model as “a simplified or idealized description or conception of a particular
system, situation, or process, often in mathematical terms, that is put forward as a basis
for theoretical or empirical understanding, or for calculations, predictions, etc.; a
conceptual or mental representation of something”. The Glossary of Geology (American
Geological Institute, 2007) defines model as “a working hypothesis or precise
simulation, by means of description, statistical data, or analogy, of a phenomenon or
process that cannot be observed directly or that is difficult to observe directly. Models
may be derived by various methods, e.g. by computer, from stereoscopic photographs, or
by scaled experiments”.
Based on the definitions of a model, predictive modeling can be defined as “making
descriptions, representations or predictions about an indirectly observable and complex
real-world system via (quantitative) analysis of relevant data”. It involves a target
variable of interest, which is usually the behaviour (e.g., presence or absence) of an
indirectly observable and complex real-world system (e.g., mineralisation), and a
number of explanatory or predictor variables or properties that are directly observable or
measurable as well as considered to be inter-related with each other and related to that
system. Predictive modeling is therefore based on (a) inter-relationships amongst
predictor variables, which may reveal patterns related to the target variable and (b)
relationships between the target and predictor variables. The latter means that some
quantity of data associated directly with the target variable must be available in order to
create and to validate a predictive model. A predictive model is a temporal snap-shot of
the system of interest, meaning that it embodies the knowledge and/or data sets used at
the time of its creation and thus it must be updated whenever new knowledge and/or
relevant data sets become available. Therefore, any cartographic representation of an
indirectly observable and complex real-world system is a predictive model.