Solving the mystery of institutions with machine learning algorithms

Nieuws - 27 februari 2018 - Webredactie

Amineh Ghorbani has managed to attract funding to do research into the DNA of institutions over a period of 700 years in the project MIDI. She is keen to study how institutions have evolved over time and will use evolutionary algorithms to find common patterns of institutional change in a unique historical dataset of institutions in Europe. It is an exceptional project in which she will work together with an international team, consisting of an evolutionary biologist, sociologist and historian. The project will result in an agent based model that will not only be used to discover patterns in history, but can also be used for contemporary problems such as the management of natural (water scarcity, deforestation) or human-made resources (community energy systems, open data).

Structuring data
Ghorbani: “How institutions evolve has always been a mystery. Many scientists have tried to solve it, but it proved to be difficult. With modern tools such as machine learning we can plough through huge amounts of data and structure this data to find answers in how our ancestors dealt with challenges such as drought for example. Which rules helped them dealing with water scarcity and how did these rules of behaviour evolve into formal institutions? Not just for one case, but several cases of commons management institutions in the Netherlands, Spain and Italy over a period of 700 years”. Many problems in history would have a close link with climatic conditions. Fortunately, the researchers also have access to environmental data, to make the picture complete. “We are looking to find common characteristics in rules and timing in these cases. This insight will be input for our model that can be used for solving today’s problems” says Ghorbani.

More information
The project has started in January 2018 and will run for two years. It is funded by the Bank of Sweden. In total 400.000 euro is awarded of which one fourth will flow to TU Delft. Amineh Ghorbani, assistant professor at the ESS department, and one research assistant will be carrying out the research.