Elia Liitiäinen’s research while affiliated with Ghent University and other places

The asymptotic variance and normality of a general class of nearest neighbor based residual variance estimators is analyzed. Based on the asymptotic results, a method for estimating confidence intervals is proposed and demonstrated via empirical simulations. The computation is easy to implement using an auxiliary estimator and a simulation technique to precompute asymptotic coefficients.

Research Question: One of the aims of this paper is to examine accounting conservatism using a robust set of data collected from the recent years to better understand how accounting conservatism affects the relations between stock returns and accounting variables. Motivation: Financial reporting is increasingly dependent on an in-depth understanding of the imperfections in the capital markets as well as the impact of the accounting standards on firm performance. Idea: The effect of accounting regulations on the firms’ reporting practices can be evaluated via a firm’s change in cash investments and its operating assets. Data: This study employs a robust set of data with different market to book ratios and corporate governance characteristics collected from Compustat North America Fundamentals Annual firm year observations for the time period of January 1998 to December 2015 fiscal years to better understand how accounting conservatism affects the relations between stock returns and accounting variables. Tools: Compared to the previous studies, any potential improvement in the research method is provided using a regression weight averaged over the years. Findings: There is an improvement in the explanatory power of the estimates of the coefficients on earnings levels and earnings changes when the variables associated with accounting conservatism are incorporated in the analysis. Contribution: Given the ample amount of research done in the international aspects of financial reporting, analyses of differences in results compared to the previous studies and future research opportunities will be provided.

We extend previously known asymptotic results in the framework of subadditive Euclidean functionals and the related superadditive functionals. Our results generalize the standard i.i.d. assumption vastly expanding the scope of the existing theory. As an application of our results, a limit theorem for the average nearest neighbor distance is established extending earlier results in the literature.

The benefits of using online exercises have been analysed in terms of distance learning, automatic assessment and self-regulated learning. In this study, we have not found a direct proportional relationship between student performance in the course exercises that use online technologies and the exam grades. We see that the average submission rate to these online exercises is not positively correlated with the exercise points. Yet, our results confirm that doing exercises along supports student learning and skill accumulation equipping them with the knowledge of programming. While the student performance in programming courses is affected by factors such as prior background in programming, cognitive skills and the quality of teaching, completing the course exercises via learning-by-doing is an indispensable part of teaching. Based on the student feedback from the course survey, the students are highly satisfied with using online technologies as part of learning.

One of the central topics in the research of higher education is the degree of match between education and work as measured by earnings. Traditional literature has shown the correlation between the duration of studies and earnings in the work life. The authors' present study aims to enlarge this scope by evaluating the impact of competence and skills learned in the university with that of the skills demanded in the work life in an empirical study. In this paper, the authors have collected elementary data from the Finnish university graduates with a respectively high response rate. Consequently, using the data of these university graduates they aim to demonstrate the mismatch between the demand for academic knowledge and the professional demand in the work life and draw policy implications on the findings. The authors will also show how these phenomena affect higher education and earnings through the analysis of skills and competence. New research opportunities are opened in the survey evidence on skills learned at the university.

This article uses data from a major international survey to construct earnings functions in terms of learning outcomes and variables related to working life in different European countries. In order to complement the extended earnings regression model, the authors have used partial correlation analysis and the analysis of covariance (ANCOVA) to discover the relevant competences appreciated by employers. Whilst one of the authors' focus countries is Finland, research results from comparison European Union countries with similar education systems have also been analysed. Consistent with previous research, the study outlines the correlation between the competences taught in universities and performance in working life, and thus quantifies the effect of higher education on earnings. In line with the theory of human capital, those who have invested more in higher education also seem to achieve higher return from the job market. However, the analysis also shows how a few of the competences taught in universities correlate negatively with the demands of working life in contrast with initial expectations. Consequently, this research offers new insights with regard to existing literature on higher education and earnings functions. The authors also open up new research avenues within the theory of human capital.

This study explores the implications of higher education on earnings in Finland. The challenges as well as opportunities of obtaining a university degree as compared to graduating from polytechnics are evaluated using the REFLEX (The Flexible Professional in the Knowledge Society) data. As a Nordic country, Finland is known for its educated population and the public emphasis on continuous learning but, nevertheless, there has been comparatively little research published internationally. The authors address the criticisms and drawbacks of higher education by analysing the performance of Finnish graduates in the labour market using the analysis of covariances. While debates concerning additional financing for higher education persist, it is acknowledged to be one of the solutions to the aging population and related economic problems in Finland. Based on the results from various Finnish educational institutions, the authors analyse the degree of match between investment in human capital and market demand.

We study the moments E[d α 1,k ] of the k-th nearest neighbor distance for independent identically distributed points in n . In the earlier litera-ture, the case α > n has been analyzed by assuming a bounded support for the underlying density. The boundedness assumption is removed by assuming the multivariate Gaussian distribution. In this case, the nearest neighbor distances show very different behavior in comparison to earlier results. In the unbounded case, it is shown that E[d α 1,k ] is asymptotically proportional to M −1 log n−1−α/2 M instead of M −α/n as in the previous literature.

In this paper, the effect of dimensionality on the supervised learning of infinitely differentiable regression functions is analyzed. By invoking the Van Trees lower bound, we prove lower bounds on the generalization error with respect to the number of samples and the dimensionality of the input space both in a linear and non-linear context. It is shown that in non-linear problems without prior knowledge, the curse of dimensionality is a serious problem. At the same time, we speculate counter-intuitively that sometimes supervised learning becomes plausible in the asymptotic limit of infinite dimensionality.

The Delaunay tessellation and topological regression is a local simplex method for multivariate calibration. The method, developed within computational geometry, has potential for applications in online analytical chemistry and process monitoring. This study proposes a novel approach to perform prediction and extrapolation using Delaunay calibration method. The main property of the proposed extension is the continuity of the estimated regression function also outside the calibration domain. To support the presentation, an application in estimating the aromatic composition in Light Cycle Oil by Near Infrared spectroscopy is discussed.