The Linear Estimation Problem and Information in Big-Data Systems
This paper addresses the problem of transforming the optimal linear estimation procedure in such a way that separate fragments of initial data are processed individually and concurrently. A representation of intermediate information is proposed that allows an algorithm to concurrently extract this information from each initial data set, combine it, and use it for estimation. It is shown that, on an information space constructed, an ordering is induced that reflects the concept of information quality.
KeywordsBig Data linear estimation canonical information distributed data collection and processing systems information algebra and information space
Unable to display preview. Download preview PDF.
- 4.Pyt’ev, Yu.P., Reduction problems in experimental investigations, Math. USSR Sb., vol. 48, no. 1, pp. 237–272.Google Scholar
- 6.Golubtsov, P.V., Informativity in the Category of Multivalued Information Transformers, Problems Inform. Transmission, vol. 34, no. 3, pp. 259–276.Google Scholar