ANTICIPATION: PREDICTIVE ANALYTICS

THE THIRD A OF THE MACROS TRIPLE A: ANTICIPATION = PREDICTIVE ANALYTICS

The opportunities businesses have due to big data, data science and machine learning are vast. However, the amount of options can blur the vision for the really meaningful measures considering industry, department and business goals. Macros Reply has worked with its experts to develop use cases that can amount to far-reaching changes for insurance companies, banks and financial service providers.

PREDICTIVE ANALYTICS USE CASES: ANTICIPATING OCCURRENCES OF EVENTS

The key method with a direct benefit for businesses evaluated by Macros Reply is Predictive Analytics. These are procedures that use analysis methods and machine learning methods based on an existing database to make predictions and intelligent decisions. This results in direct benefits for companies and / or customers, as the following use cases show:

• WEATHER FORECAST:
  Use case: precise (3 x 3km) forecast of severe weather based on partner and inventory data and weather reports
  Objective: anticipate extent of damage and point in time, plan processing capacities, facilitate faster regulation
• FRAUD DETECTION:
  Use case: detection of fraud cases based on patterns, anomalies and accumulations of data
  Objective: find fraud attempts and increase the detection rate, despite fewer reference cases
• PREMIUM CALCULATION:
  Use case: determination of parameters which significantly influence the premium contribution, e.g. through tree algorithms with a very good prediction accuracy compared to traditional models (e.g. typically 10-20% better accuracy than GLM)
  Objective: more accurate premiums, risk reduction
• TREND DETERMINATION FOR BENEFIT PAYMENTS:
  Use case: early detection of anomalies, i.e. changes in benefit payments
  Objective: detection of price increases in medical services / drugs, detection of up to this point unknown cost drivers, determination of the influence of detected anomalies on contribution rates
• RISK ESTIMATION AND HEALTH PREDICTION:
  Use case: risk calculation based on different data sources (such as billing data, analysis of patient data, text mining on health questionnaires) through text mining and supervised learning
  Objective: predict and deduce state of health for active health management and to avoid benefit payments.

MACROS REPLY CONSULTING APPROACH FOR MACHINE LEARNING & PREDICTIVE ANALYTICS

• Evaluation of potential approaches to Machine Learning (possibly based on the knowledge gained from Process Mining and Robotics)
• Definition of use cases and your objectives
• Smart Start:
  1-day workshop: core requirements of the customer
  Use case evaluation for Smart Proof preparation
• Smart Proof:
  2-5 days Proof of Value
  Cooperation of experts from Macros Reply and customer
  Creating initial project proposals
• Transfer of the successful use cases to productive operation
• Monitoring the measures
• If desired, training and further education of the employees and consultation on the formation of a specialist team (Data Incubator Reply is a training program for data scientists who - building on their master's or doctor's degree - specialize in Machine Learning and Data Science within eight weeks. Companies can invite candidates or let their own staff acquire the relevant skills in this course.)

Fore more information download the flyer Macros Triple A.