In the mid-80s began Henrik begun his undergraduate studies at DSV and when he had finished, he was accepted as a doctoral student in the field of artificial intelligence, with an emphasis on machine learning. Henry was early and was lucky to meet many of the foremost in the field in the beginning of his research career.

“Much has happened during the past twenty years. The area has gone from being a research area of mainly academic interest to deliver algorithms, methods and software to all kinds of application areas, from life sciences to analysis of purchasing patterns,” Henrik says.

Today huge amounts of data in many different activities are generated, and there is great potential in using these data to improve operations. That can be predicting whether a customer might be interested in a certain product or to making predictions about the lifespan of components in vehicles used for maintenance planning.

Development in the field has means that great opportunities both to collect and to analyze data on a large scale. Companies and organizations recognize the benefits of acting on data rather than on the basis of speculation. That increases the interest in the area, which is popularly called "big data" - or data science which is the name Henrik prefers.

"Even ten years ago it felt like the market was not quite ready, but now interest has literally exploded. Now it's really hot," Henrik emphazises. "This manifests itself not least through the the great impact on the master's programs and research centers worldwide."

Research in the area at DSV focuses on data and text mining, where algorithms from machine learning are used and developed in order to make predictions based on past observations and to find patterns in the data in the form of interpretable rules. Ten senior researchers working in the field at the institution along with 12 graduate students.

“In addition to data and text mining the area of data science involves several other activities at DSV, such as software development, database technology and visualization. This is also a strong area for the Faculty of Social Sciences with links to, among other things, statistics and business administration,” Henrik explains.

Drug Effects
Project Data analysis for detection of drug effects (Dadel) has received 19 million over five years under the SSF programme information-intensive systems. The main goal of the project is to develop techniques and tools to support decision making and detection of drug effects by analyzing patient records, drug records, adverse event reports and experimental results. The project develops methods for analysis of both structured and unstructured data, and software for large-scale analysis of massive, heterogeneous and ever-growing amounts of data.

"The amount of data in the form of electronic health records, drug records, adverse event reports and experimental results is growing almost exponentially. By linking different sources of information new information can be obtained regarding the effects that drugs have," Henrik tells.

The information can support decision-making at both the development and at the prescription of drugs. However, it is very challenging to analyze these kinds of massive, heterogeneous and ever-growing amounts of data. In addition to the six senior researchers six doctoral students are part of the project, which is a collaboration with the University of Borås. This fall, there will be half-time reporting - something that he looks forward to with confidence.

"In terms of research output we are well on our way - during the first half of the project, we have already published more than 40 publications - conference papers and journal articles. Some of the articles have been published in the most prominent journals (highest level according to the rating used at the university),” Henrik emphasizes. “We are also working on demonstrators that we should be able to show at the meeting for mid-term review in November.”

The DADEL project involves exciting collaborations with researchers in pharmacology and drug development from both research institutions and pharmaceutical companies.

“We will develop predictive models and be able to demonstrate confidence-based prediction of adverse drug reactions. It is also about creating tools that make it easy for users to analyze their data to draw conclusions,” Henrik explains.

SSF wants the research they fund also to be commercialized. And it is of course very good, although Henrik does not see himself as the natural entrepreneur.

“It's very exciting to see the results of your research being used in practice. Personally, however, I prefer doing research and of creating artifacts. I would like to see collaboration with entrepreneurs to start up a commercial operation based on the results,” Henrik stresses.

Other areas of application
Results of the project are generalizable and can be applied to other areas than drug effects. In the project IRIS DSV cooperates with Scania and Linköping University. This project will assess the health status of heavy vehicles. Scania and Vinnova are investing about 23 million in total on the five year project.

To develop the Swedish transport sector there is a demand on delivery performance, security and automation. The cost of unscheduled maintenance must be reduced and there is a need about the vehicle's health status. The project will develop prediction models to be used to dynamically optimize planned maintenance.

And the future?
Henrik has always kept one foot at DSV although he had several other assignments outside the department. He has worked as a senior researcher at the startup company Virtual Genetics at DSV Karolinska Institutet Campus where he led the development of their product for predictive modeling. He also was co-founder of the company Compumine AB, that for almost ten years delivered its own product Rule Discovery System (RDS), mainly to the pharmaceutical industry. Henrik has also be been a professor of information fusion at the University of Skövde for some years, and had administrative assignments, including as deputy head of the DSV. But research he is what he is passionate about.

Henrik sees a bright future, and he hopes that the activities in data science at the department will be conducted in at least at the same scale as now. Interest in the area is large on a national level, and in early December, a workshop in data science will be held at DSV with seventy participants from universities, institutes and companies.

“The idea is that we not only learn more about each other's ongoing research project, but also discuss forms of collaboration on a larger scale, for example within the framework of a research school and joint projects,” Henrik explains .

“It's an exciting time right now, and I very happy with my life. It's also great that so many talented doctoral students are coming on to the area,” Henrik concludes

Private
Wife two daughters (soon 18 years both)
Family farm at Söderslätt in Skåne
Passionate golfers since four years (Hcp 15). Should be singelhandikappare within five years.

Personal webpage

Project participants DADEL
Henrik Boström (Project Manager), Professor, Stockholm University
Hercules Dalianis, Professor, Stockholm University
Lars Asker, Associate Professor, University of Stockholm
Ulf Johansson, Associate Professor, University of Borås
Håkan Sundell, Associate Professor, University of Borås