Transcript Presentation of the centre by Frigessi.

Arnoldo Frigessi
Department of Biostatistics
University of Oslo
[email protected]
with the support of
Oil & Energy
Markets
Smerud
Medical Research
Petroleum
Finance
Insurance
Health
Marine
Common challenges to statistical science
Common challenges to statistical science
much data from different sources, but often few critical ones
more complex data, because we measure more in depth
complex systems to be modeled, with structured dependencies
the easy part is done, necessary to discover “second order effects”
evidence-based decision making, risk evaluation is required
computationally intensive procedures
Statistical methods
We shall produce the next generation of
statistical methodology for innovation
Cutting-edge novel statistics
Cutting-edge innovation projects
Partners
•
Based on our scientific competence, experience, traditions …
• we shall built strong, interdisciplinary, ambitious teams, …
• planning together and producing top research results, …
• fundamentally needed to solve strategic innovation needs.
statistics for innovation:
The largest European research centre
in applied statistics.
for the first time in the world
Statistics focuses on Innovation!
A unique task force
“This research group ought to be regarded as a national treasure.
Many much larger countries with much greater resources should be envious of
Norway for having such a team.”
“The strengths are the high quality staff and excellent industrial commitment.
This grant should secure the position of Norway as one of the leading countries
in Europe in industrial statistics.”
Data integration - Complex models
Long Predict - Tail Depend - ByCatch
EventAnalysis - NetAlgebra - TotalRisk
StatInsure - Commodity - SampleFish
Multicompare - MissingStripes - FindOil
GenomeAtlas - MetaAnalysis
TotalRisk
• Finalise the development of a new method to estimate the total risk and the
needed total capital reserves for a risk-diversified bank/financial corporation.
Correlations
Dependencies
Credit
risk
Marked
risk
+
Experts’
opinions
Operational
risk
OSLO TOTAL
RISK MODEL
Data
MetaAnalysis
• More and more clinical trials are available, assessing the efficacy of drugs.
• How to put together conclusions across published studies? A meta-analysis
combines results of many clinical studies.
• How to merge published clinical studies with own precious data at patient level?
We will invent new methods to merge public and private data.
LongPredict
 Long-term predictions are steering instrument. Financial time series, climate
or fish abundance - there is a need to project current trends ahead in the
future, maybe 5 years, possibly 20, 50!
 The question is not predicting the detailed behaviour but particular structural
summaries: the probability that a trend is maintained, that certain (low) levels
will not be reached anymore, that a new stationary regime has been reached.
 We shall develop new
methods to answer these
questions.
Index value,
Jan 02=100
400
350
300
250
200
150
100
50
0
1990
1992
Brent Oil
1994
1996
NWE Coal
1998
Aluminium
2000
2002
Steel HRC
2004
2006
FED US$ index
FindOil
• Understanding Exploration: where to drill next
to find oil and gas?
Global
Region
Basin
Province
Play
Prospect
• Develop a new methodology to guide exploration,
by computing more precise probabilities of
success.
SampleFish
 Fish quotas are made on estimates of abundance of fish, based on data from
commercial fishery, and sampling survey data.
 We shall improve the sampling, develop new profound mathematical models that
describe the marine environment under environmental pressure, climatic change and
commercial exploitation.
 This will lead to a new standard for a affordable management of marine resources.
International alliances.
Building on our academic and corporate international contacts, we will create a
truly international centre, promoting international projects and alliances that
build bridges between academia, business and industry.
Training and education.
Industry-based research projects for master, PhD students, postdocs.
UiO international graduate training programme in (bio)statistics for innovation
… with the aim to enhance the
capability of the Norwegian economy
to innovate by focusing on long-term
research, forging alliances between
research-intensive enterprises and
prominent research groups.
A broad view of the world - the global role and responsibility of science & technology.
Much of the improvement in human welfare over the past can be accounted for by
technological innovation in energy, engineering, public health, nutrition, fishery,
agriculture. Statistics played a role also here.
is aware of this role. Many partners are
active with training, education, projects and
investments in the developing parts of the world.
KLAR TIL
START!
Takk til
Norges Forskningsråd!
Arnoldo Frigessi
[email protected]