510 and ORTEC are collaborating to offer two master Students to be part of a unique organisation and team in the field of Text mining. The two internships will inform both how text mining can help with regards to Vulnerability and Capacity Assessment & Disaster Relief Emergency Fund.In the most general terms, text mining “turns text into numbers” (meaningful indices), which can then be incorporated in other analyses such as predictive data mining projects, the application of unsupervised learning methods (clustering), etc.
WHY TEXT MINING & DREF
Every year, many small and medium-sized disasters occur in silence, without the attention of the mainstream media. To support these smaller emergencies or disasters, or to provide initial funding before emergency appeals are launched for large-scale disasters, the International Federation of Red Cross secretariat allocates grants from its DREF to National Societies to support their operations. All requests for DREF allocations are reviewed on a case-by-case basis. In relation to the DREF the following documents are produced: appeals, plans and updates. These documents are available through the following repository:
The aim is to widen the application scope of DREFs documents for disaster response and preparedness in order to better shape and target future interventions. Data- and specifically text mining should be employed to semi-automatically extract data from DREF reports, by analysing ways to combine this data with open, public risk data and provide recommendations as to how to structure future digital DREFs.
WHY TEXT MINING & VCA
VCA is a process developed within the Red Cross and Red Crescent Movement. It supports communities to become more resilient through the identification, assessment and analysis of the risks they face. The VCA Enhancement Process was launched in 2016 to improve understanding, quality information sharing, capacity and coordination around the VCA, see https://www.ifrcvca.org/. This process resulted in a large repository of VCA-reports which probably contain a lot of value, however in a very ‘unstructured’ and thus not actionable manner.
The aim is to widen the application scope of VCAs to incorporate disaster risk reduction, climate change adaptation and disaster preparedness. Data- and specifically text mining should be employed to semi-automatically extract relevant risk data from historical VCA reports. It is also interesting to analyze ways to combine this data with open, public risk data and to come up with recommendations as to how to structure future digital VCAs so that the incorporation of data and text mining becomes straightforward.
So as a student you will help on increasing the impact of the VCA.
During the course of a 6/7 Months thesis, both organizations will work with the graduate students on developing ways to use text mining to benefit VCA and FUNDS.
Dr Marc van den Homberg, the scientific lead of 510
Ronald Buitenhek, the lead of the Center of Excellence Machine Learning at ORTEC.