Final report Tools for information fusion and analysis

This report provides a summary of the work within the project Tools for information fusion and analysis, which started in 2013 and ended in 2015. The aim of the project has been to build and transfer knowledge to the Swedish Armed Forces about possibilities and limitations with data driven intelligence analysis. In year one, a number of research needs at the Armed Forces were identified, leading to a focus for the project on assessing the trustworthiness of information on the Internet, with a special interest in deception. A theoretical work was undertaken in order to identify various aspects of deception on Twitter and find out how such deception can be detected. In 2014, development of a platform for experimenting with new methods and techniques to support analysts with deception detection was initiated. The first step towards such a platform was to implement and deploy a retrieval module for creating datasets to make experiments on. Furthermore, three development tracks were initiated, one focusing on visual analytics, one on natural language processing, and one on data mining. As the developed tools started to mature, analysts were engaged to give feedback and suggestions at an early stage. During 2015, the development of the tools has continued and they have been integrated into a common prototype system for detection of influence operations on Twitter. This system has been transferred to a Hadoop cluster in order to work for analysis of big data. A final demo has been carried out with the Swedish Armed Forces and related stakeholders.