Generally, predictive modelling refers to “scoring” data with predictive analytics and forecasting techniques. However, the term “predictive analytics” is used to refer to various disciplines/models, such as descriptive models, predictive models or decision models (“Predictive Analytics”, 2015).
Predictive models analyse the relation between one or more attributes of the input data (“Predictive Analytics”, 2015). These models are used to estimate the likelihood that a similar unit will exhibit the same performance in different sample. This category focusses on modelling in many fields, such as operational management, where they understand the complex data patterns revealing the customer performance and identifying fraudulent activities in a business. Predictive models perform calculation over live transactions, in order to solve many issues related to operational management, such as to evaluate the risk to the organization.
The sample units whose attributes and performances are known, is known as “training samples”. While other sample units whose attributes are known but performance is unknown, is called “out of sample” units. The “out of sample” units do not have any relation with “training sample” units. For example, during a simulated crime scene, the blood splatter pattern is an example of “out of sample” units which needs to be analysed.