Spectral preconditioning of Krylov spaces: Combining PLS and PC regression

In regularized regression the vectors that lie in Krylov and eigen subspaces, formed in PLS and PC regressions respectively, provide useful low dimensional approximations for the LS regression coefficient vector. By preconditioning the LS normal equations we provide a framework in which to combine these methods, and so exploit both of their respective advantages. The link between the proposed method to orthogonal signal correction and to cyclic subspace regression is made. The performance of the proposed solution in reducing the dimension of the regression problem, and the shrinkage properties of the resulting coefficient vector, are both examined.     

The influence of ICT-driven innovation: a comparative study on national innovation efficiency between developed and emerging countries

ABSTRACT

In the Big Data era, Information and Communication Technology (ICT), including the Internet and sensors, digitises physical activity extensively. This leads to the development of ICT Driven Innovation (IDI) which may have a strong influence on National Innovation Efficiency (NIE). The purpose of this paper is to provide insights into the impact of IDI on NIE in both advanced and emerging countries. Data Envelopment Analysis (DEA) is employed to obtain the individual score for each country. We focus on comparing IDI between advanced and emerging economies in particular. We believe that IDI is an important reason why the average NIE score is higher in developed countries than in emerging countries. Obviously, developed countries have shown good capabilities in IDI in excess of most emerging countries. This includes but is not limited to R&D expenditure across the ICT industry, ICT patents etc. Several emerging economies received high NIE scores such as China and Brazil, who also have a good performance in IDI. On this basis, we discuss the mechanism of how IDI affects NIE. ICT industry innovation, non-ICT industries innovation, and ICT infrastructure are summarised as the primary factor affecting NIE by IDI. Further empirical research is required in the future.