A guest post by Xi Shi from KU Leuven
Nowadays, people’s working style has been changed disruptively by the rapid development of Internet technology. With the establishment of electronic administrative system, the governments gradually accumulate abundant data from various data sources. However, in most cases, the data is stored in the database without effective utilization. There is a growing need to make full use of the data.
Simulation is one of the effective techniques to imitate and approximate the operation of a process or a system [1]. It can show the eventual impacts or influences of alternative conditions and courses of action. Simulation is often applied when the real system cannot be engaged, because it may not be accessible, or it may be dangerous or unacceptable to engage, or it is being designed but not yet built, or it may simply not exist [2]. Because of these properties, simulation plays an important role in decision making. Simulation makes it possible to compare a proposed process with the existing one, to make changes and see the effects in real life without any risk to the current system. If the decision is complex, simulation can help to assess the impact of changes on every aspect of the process. Forecasting is another application of simulation. When the simulation model is constructed and the decision making process is well approximated, we assume all the conditions and circumstances remain constant or change at a defined pace, then we can use the simulation model to predict future trends and developments.
MIDAS project integrates resources from universities, companies, and governments to develop a platform to provide data analytics and visualization for better policy decision making. In the context of MIDAS project, our work is to provide the basis to derive insight to support policy decision making with the use of scalable data analytics. This will influence policy development through an extended process of communication and interactions between policy makers. To support health policy decision makers, we envision tools that allow them to simulate the effect of certain policy changes on various aspects of health care, including clinical outcomes, demand, quality of life, and public health economics. These tools are also of great value for planning service implementation and delivery. For this purpose, we will rely on proven methods from public health, demography, econometrics, and forecasting, and adapt them to the specific use case when necessary. Once a list of key policy questions to tackle is established, we will study which statistical and machine learning algorithms are required to extract the relevant knowledge from the data at hand to offer decision support to relevant policy makers. We envision the use of characteristic methods of all main forms of statistical learning, i.e., regression, classification, clustering, density estimation, and time series analysis.
References
J. Banks; J. Carson; B. Nelson; D. Nicol (2001). Discrete-Event System Simulation. Prentice Hall. p. 3.
J. Sokolowski; C. Banks (2009). Principles of Modeling and Simulation. Hoboken, NJ: Wiley. p. 6