Transcript JDemetra+
SASG, 16/10/2012
What's new?
◦
◦
◦
◦
◦
Graphical interface
Statistical issues
I/O features
Technology
Tests
Final remarks
Same main principles as .NET version
• Interactive processing
• Rich results
• ...
Many improvements and goodies
See the presentation
Core engines (Java implementation)
◦ X12/X13:
Small improvements in comparison with the previous
Java version
◦ Tramo-Seats:
≈ future release of Tramo-Seats (FORTRAN):
New AMI procedure
New strategy in the case of non decomposable models
Automatic choice of the "best" TD/WD effects (in progress)
! Stochastic TD not integrated !
Other SA methods
◦ Structural models, Generalized airline, Airline + seasonal
noise
Benchmarking
◦ Univariate Cholette's method (see X13)
◦ Extension to the multi-variate case with contemporaneous
and/or temporal constraints
Seasonality tests
◦ See Tramo
Direct/indirect comparison
◦ Use of multi-variate benchmarking if need be.
Calendars
◦ Weights on holidays
Improved access to external data
◦ ODBC, Spreadsheet (Excel or OpenOffice)...
More output (larger set of results)
◦
◦
◦
◦
Csv
Csv matrix
Excel
Txt
JTsToolkit
Core algorithms
External
packages
Peripheral
modules
In house
developments
Pure Java
NetBeans modules
NetBeans
JDemetra+ plugins
Third party
plug-ins
JDemetra+ = NetBeans application
What is NetBeans ?
IDE for Java developments (and others)
Framework for extensible applications
Sponsored by Oracle
Benefits of using NetBeans?
◦ Extensible architecture
Allows independent development teams
New features = new plug-ins (no impact on the existing
modules)
◦ Numerous functionalities
Rich graphical interface
Management of the plug-ins
Automatic updates...
◦ (Well) documented framework
◦ Large developers' community
◦ So, JDemetra+ = { plug-ins}
Core SA (TramoSeats, X13...)
Advanced SA
Benchmarking
...
Data providers
...
Profound re-engineering of Demetra+ (.NET)
◦ Better design of the algorithmic modules
Extensible algorithms
Faster processing (huge use of multi-processing)
◦ Improved I/O of the high-level components
Generic xml serialization
Facilities for providing results to other environments
Designed for
(Quasi-)immediate WEB services
Development of macro-languages (?)
◦ New design of the other modules to fit higher
(NetBeans) modularity
Direct calls to the algorithmic routines
◦ Should be the preferred solution in many cases.
◦ From simple high level modules...
◦ ...to all details
◦ Depends only on 1 library (jtstoolkit.jar)
◦ Documentation based on examples (in progress)
◦ Training in November
Contents of the JTsToolkit API (≈50% code)
◦ Basic mathematical tools
Matrices, polynomials, filters, optimization, ...
◦ Advanced time series model
◦ Statistical tools
Descriptive statistics, statistical distributions, ...,
state-space framework
Designed to develop rapidly new features
◦ For examples:
New tests (Canova-Hansen...)
Time-varying TD.
Temporal disaggregation (largely developed)
...
Must be NetBeans modules (-> for advanced
users/developers)
Possible extensions (open list)
◦ Modification of the existing features (graphical
components, menus...)
◦ New data providers, new diagnostics, new output
◦ New seasonal adjustment methods (automatic
integration in batch processing...)
◦ New statistical methods (integration in the main
menus, in the workspace...)
Such extensions will be installed as plug-ins, without
modification of the existing modules.
Needs for test
◦ validation of the statistical methods
Type of algorithms
◦ "Fuzzy" problems:
Example: seasonal adjustment methods...
Difficult to develop a true strategy.
◦ " Implicit" problems:
Example: non linear optimization (ML estimation)...
Possibility to compare the results. No way to guarantee an
optimal algorithm
◦ "Analytical" problems:
Example: Estimation of a Reg-Arima model...
Possibility to validate an algorithm
Main tests developed in JDemetra (type 3)
◦ Likelihood computation of Reg-Arima models
Comparison of Kalman filter (Tramo), Ljung-Box algorithm( X13)
and others (Ansley...).
◦ Solution of linear models
Comparison of QR, LU, singular value decomposition...
◦ Canonical decomposition
Verification through decomposition/aggregation
◦ Many small other tests
Benchmarking, statistical distributions, auto-correlations, linear
filters, Easter...
Verification by:
comparison of alternative algorithms
logical tests (checking of some constraints...)
tests by reciprocal transformation
◦ Further tests needed
Huge developments supported by NBB and
Eurostat
NBB will ensure a minimal service of
maintenance:
◦ Correction of bugs
◦ Training
◦ Add-ins following the needs and the agenda of NBB
Necessity to organize the future