Descriptive analytics or unsupervised learning aims at finding unusual anomalous behavior deviating from the average behavior or norm (Bolton and Hand 2002). This norm can be defined in various ways. It can be defined as the behavior of the average customer at a snapshot in time, or as the average behavior of a given customer across a particular time period, or as a combination of both. Predictive analytics or supervised learning, as will be discussed in the following chapter, assumes the availability of a historical data set with known fraudulent transactions. The analytical models built can thus only detect fraud patterns as they occurred in the past. Consequently, it will be impossible to detect previously unknown fraud. Predictive analytics can however also be useful to help explain the anomalies found by descriptive analytics, as we will discuss later.
Azure Claim Fraud Analytics Solution leverages Azure Synapse unified platform in bringing together different aspects of data like data ingestion, big data analytics and data warehousing. Health Insurance users like Data Architects, Data Scientists, Data Analysts and Business users are facilitated by this solution. Azure Synapsis claim fraud analytics solution uses descriptive and predictive modelling with industry standard business rules to identify, score, and prioritize possible fraud cases.
The overwhelming majority of executives say that their organisation has achieved successful outcomes from Big Data and AI. Data can also have a big impact on your bottom line, with businesses who utilise big data increasing their profits by an average of 8-10%. Netflix reportedly saves $1 billion every year by using data analytics to improve its customer retention strategies.
Descriptive analytics are often displayed using visual data representations like line, bar and pie charts and, although they give useful insights on its own, often act as a foundation for future analysis. Because descriptive analytics uses fairly simple analysis techniques, any findings should be easy for the wider business audience to understand.
These predictions can then be used to solve problems and identify opportunities for growth. For example, organisations are using predictive analytics to prevent fraud by looking for patterns in criminal behaviour, optimising their marketing campaigns by spotting opportunities for cross selling and reducing risk by using past behaviours to predict which customers are most likely to default on payments.
Another branch of predictive analytics is deep learning, which mimics human decision-making processes to make even more sophisticated predictions. For example, through using multiple levels of social and environmental analysis, deep learning is being used to more accurately predict credit scores and, in the medical field, it is being used to sort digital medical images such as MRI scans and X-rays to provide an automated prediction for doctors to use in diagnosing patients.
Businesses that can harness the power of prescriptive analytics are using them in a variety of ways. For example, prescriptive analytics allow healthcare decision-makers to optimise business outcomes by recommending the best course of action for patients and providers. They also enable financial companies to know how much to reduce the cost of a product to attract new customers whilst keeping profits high.
Predictive modeling is but one aspect in the larger predictive analytics process cycle. This includes collecting, transforming, cleaning and modeling data using independent variables, and then reiterating if the model does not quite fit the problem to be addressed.
Similarly, with marketing analytics, predictive models might use data sets based on a consumer's salary, spending habits and demographics. Different data and modeling will be used for banking and insurance to help determine credit ratings and identify fraudulent activities.
Modern analytics tend to fall in four distinct categories: descriptive, diagnostic, predictive, and prescriptive. How do you know which kind of analytics you should use, when you should use it, and why?
Predictive analytics is a form of advanced analytics that determines what is likely to happen based on historical data using machine learning. Historical data that comprises the bulk of descriptive and diagnostic analytics is used as the basis of building predictive analytics models. Predictive analytics helps companies address use cases such as:
Prescriptive analytics is the fourth, and final pillar of modern analytics. Prescriptive analytics pertains to true guided analytics where your analytics is prescribing or guiding you toward a specific action to take. It is effectively the merging of descriptive, diagnostic, and predictive analytics to drive decision making. Existing scenarios or conditions (think your current fleet of freight trains) and the ramifications of a decision or occurrence (parts breakdown on the freight trains) are applied to create a guided decision or action for the user to take (proactively buy more parts for preventative maintenance).
Prescriptive analytics requires strong competencies in descriptive, diagnostic, and predictive analytics which is why it tends to be found in highly specialized industries (oil and gas, clinical healthcare, finance, and insurance to name a few) where use cases are well defined. Prescriptive analytics help to address use cases such as: 59ce067264