Page 103 - Työpoliittinen aikakauskirja 1 2018
P. 103

Työpoliittinen aikakauskirja 1/2018
English Summaries
  Algorithm to improve job-worker matching
Jaakko Matomäki, M.Sc., Data Scientist; Tiia Leuhu, M.Sc., Data Scientist; Jouni Alin, M.Sc., Analytics Leader, Deloitte Consulting Oy (all authors)
This study aims to develop an algorithm to match jobs and workers in the Finnish job market. The  rst part of the study is a literature review of the state of the art methodologies in job matching. The study used data from Finnish job seekers and open jobs between 2015 and 2017 from URA system. Job o ers leading to employment were de ned as good matches.
The performance of seven di erent algo- rithms with di ering hyperparameters and miss- ing data strategies were evaluated. In total more than 100 combinations were evaluated in their ability to identify suitable jobs for job seekers. A random forest algorithm outperformed other algorithms. Location and work history were the most important predictor variables for good matches.
Advantages and disadvantages of the solution were evaluated. In addition, the study identi ed possible external data sources that could be used to further improve the matching.
tion is not consistent with simple skill-bias tech- nological change hypothesis. New approaches highlight the role of tasks performed in di erent occupations or jobs as mediators of the e ects of technological change and o shoring to the wage and employment structure. The objective of this article is to examine if, and how, the changes in the occupational wage and employment patterns relate to the di erences in tasks performed in di erent jobs in the Finnish setting. To accom- plish this, we exploit a new set of task measures speci cally developed for the European labor markets by the Eurofound. These measures build on the classi cation of job tasks along two main dimensions: the content of the tasks themselves (what people do at work) and the methods and tools used to perform those tasks (how peo- ple work). We link these task measures to the Harmonized Wage Structure Statistics compiled by Statistics Finland. It is individual-level data for the private sector of the Finnish economy covering the period 1995-2013. Using this linked data, we analyze how changes in occupational wages and employment relate to the task con- tent of occupations in Finland over the period 1995-2013.
Robots are gradually being integrated to wel- fare services and to “soft” environments such as the elder care. At the same time robots are being developed as more suitable for human interac- tion. Care robots should have arti cial intelli- gence to the point of even simulating empathy. Also care robots’ exterior is under examination. Robots are being developed as more human-like with  ne motor skills and even fake skin. Care robots have potential to assist nurses in their work or directly, for example, home care custom-
 Do Robots take over care work?
Tuuli Turja, BA Psych, M.Soc.Sc, Researcher – Tuomo Särkikoski, Msc, Lic.Soc.Sci, PhD, Senior Researcher, University of Tampere (both authors)
 Tasks and changes in occupational wages and employment in Finland
Jari Vainiomäki, PhD, professor, University of Tampere
Petri Böckerman, PhD, professor, University of Jyväskylä and Labour Institute for Economic Research
The traditional approach to examine the changes in wage di erences and employment structure focuses on worker skills and their returns to e.g. education and work experience, or labor market institutions. Recent studies have stressed that the concentration of both wage and employment growth to the extreme ends of the wage distribu-

   101   102   103   104   105