Deliverables Methods for handling ill-conditioned data in the NIS The main activities were a write up of the respective ideas and models for handling missing data of the NIS members, learning the western MI methods and their implementation by NIS participants. First results and initiatives were made public at four international conferences in Russia, the Slovak Republic, USA, Austria, Spain and the Ukraine. The next activities were the follow-up of a workshop in Leiden on 5 July 1996. And of the study trips by three NIS participants and the twinning of two NIS junior scientists for three months at TNO PG in Leiden, The Netherlands. The simulation of the TNO missing data machine (MICE) on Russian grounds through CHILD was the main achievement. A NIS proposal to generate the problem-oriented metrics for genetic sequences, with possible application for algebraic methods in data conditioning was developed. Main results of the activities are the introduction of the applied use of empirical Bayes methods (including the Gibbs sampler) for handling missing data to the NIS statistical community. This is done by letting the NIS participants do the actual building of the implementation and the actual conceptualising involved. The area of application within applied statistics is missing data. This in itself is a relatively new subject for the NIS statistical community. Tangible products are: 15 Presentations on this new Western technology in Russian, in Russia and Ukraine by NIS participants. The establishment of a working centre for handling Ill conditioned data (CHILD) in Moscow. A monograph in English on the subject to be published in 1999 by CCAS publishers Moscow. Searching for OpenAIRE data... There was an error trying to search data from OpenAIRE No results available
