Predictive Analytics
Overview
The work with time series can be divided into three areas. These areas are called Descriptive Analytics, Predictive Analytics, and Prescriptive Analytics and can be distinguished as follows:
- Descriptive Analytics analyzes and describes historical data, identifies patterns, highlights relationships, or detects anomalies in the data.
- Predictive Analytics builds upon this by making statements about the future, generating data-driven forecasts, and estimating the probabilities of potential future events.
- Prescriptive Analytics translates future predictions into concrete recommendations for action and identifies optimal decisions.
Definition
Predictive Analytics identifies patterns, structures, and relationships in data and uses them to make statements about future events. It employs methods from classical statistics as well as machine learning.
Areas of Application
Predictive Analytics has a wide range of applications.
- Supply Chain – Intelligent supply chain, production, and inventory management: Demand for certain items in a company is forecasted using order history, digital customer information, and current order inputs. Accurate predictions form the basis for optimizing and coordinating raw material purchases, production, and inventory management.
- Smart Grid – The intelligent power grid: The smart grid generates load forecasts and predicts electricity demand. Such load forecasts are becoming increasingly important due to the decentralization of power generation (renewable energies). This is relevant for both the power provider, who must ensure grid stability, and the energy-consuming company, which needs to adapt to the volatile energy market as cost-effectively as possible.
- Predictive Maintenance – Proactive maintenance planning: Intelligent production facilities detect wear and malfunction of machines early using sensor data, allowing maintenance activities to be scheduled on time or coordinated with other planned maintenance tasks.
- Credit Scoring – Probability of loan payment defaults: In credit scoring, banks and companies assess the likelihood that a customer will fail to meet installment payments on a granted loan, enabling them to adjust the terms of the loan accordingly.