Data-Driven Secure Business Intelligence
- Reference number
- IIS11-0089
- Start and end dates
- 120101-171231
- Amount granted
- 24 987 385 SEK
- Administrative organization
- Chalmers University of Technology
- Research area
- Information, Communication and Systems Technology
Summary
Develop scalable system architectures, algorithms, development methods, and working demonstrators for temporal analysis of large data sets harvested from open sources (web, social media etc.) as well as corporate databases (customer data, business intelligence data) to enable new forms of collaborative innovation. These analysis services must scale to handle very large data sets, and have mechanisms for ensuring privacy and personal integrity for individuals as well as security for customers. Applications include competitive business intelligence, continuous product development, predictive analytics and many other areas of great importance to Swedish industry, both as providers and users of these services. Concrete demonstrators include: - Predictions of financial markets (Recorded Future, First Swedish Research) - Analysis of consumer behaviour and predictions about future behaviour respecting privacy (RF, TeliaSonera) - Service development through experimental and collaborative innovation (RF, Tibco Spotfire) The work will proceed in an iterative fashion with implementation and testing of methods leading to feedback into disciplinary research for improved methods which are then implemented and tested again. In the first two years we will develop prototypes to test our methods on the Recorded Future database and in the following years we will scale them up to integrate into the RF system implemented on the Amazon EC2 cloud architecture.
Popular science description
There has been a convergence of information technologies and social processes leading to what is called Web 2.0 or the social web. As a result, companies, Governments and individual users are connected ever more tightly in a two-way interaction that has generated an unprecedented explosion of data. This data can be used to understand social behaviour, customer preferences, business trends etc. and generate predictions for the future. In this way risks and waste can be ameliorated, business efficiency can be improved and better and newer value can be provided to customers. Machine learning is an area of Computer Science that aims to develop techniques to extract meaningful information out of huge masses of data. This project aims to apply these techniques to pressing business and social needs and to develop them further. When it comes to data about individuals, privacy is huge public concern. Even with first-class security measures in place it may be possible to extract far more information about an individual than would be considered acceptable. We plan to use and extend the state of the art research, differential privacy, for extracting useful data without compromising the privacy of individuals - applied for the first time in an industrial context. We will work closely with new small and medium sized Swedish firms (Recorded Future, Tibco Spotfire and Swedish First Research) and a large supplier of connectivity services (TeliaSonera) to demonstrate the benefits of this research. A concrete demonstrator is targeted to the financial domain and to provide tools for analysts to safeguard individual privacy.