Statistical analysis of survey-based event history data with application to modeling of unemployment duration

Longitudinal surveys are increasingly used to collect event history data on person-specific processes such as transitions between labour market states. Surveybased event history data pose a number of challenges for statistical analysis. These challenges include survey errors due to sampling, non-response, attrition and measurement. This study deals with non-response, attrition and measurement errors in event history data and the bias caused by them in event history analysis. The study also discusses some choices faced by a researcher using longitudinal survey data for event history analysis and demonstrates their effects. These choices include, whether a design-based or a model-based approach is taken, which subset of data to use and, if a design-based approach is taken, which weights to use. The study takes advantage of the possibility to use combined longitudinal survey register data. The Finnish subset of European Community Household Panel (FI ECHP) survey for waves 1–5 were linked at person-level with longitudinal register data. Unemployment spells were used as study variables of interest. Lastly, a simulation study was conducted in order to assess the statistical properties of the Inverse Probability of Censoring Weighting (IPCW) method in a survey data context. The study shows how combined longitudinal survey register data can be used to analyse and compare the non-response and attrition processes, test the missingness mechanism type and estimate the size of bias due to non-response and attrition. In our empirical analysis, initial non-response turned out to be a more important source of bias than attrition. Reported unemployment spells were subject to seam effects, omissions, and, to a lesser extent, overreporting. The use of proxy interviews tended to cause spell omissions. An often-ignored phenomenon classification error in reported spell outcomes, was also found in the data. Neither the Missing At Random (MAR) assumption about non-response and attrition mechanisms, nor the classical assumptions about measurement errors, turned out to be valid. Both measurement errors in spell durations and spell outcomes were found to cause bias in estimates from event history models. Low measurement accuracy affected the estimates of baseline hazard most. The design-based estimates based on data from respondents to all waves of interest and weighted by the last wave weights displayed the largest bias. Using all the available data, including the spells by attriters until the time of attrition, helped to reduce attrition bias. Lastly, the simulation study showed that the IPCW correction to design weights reduces bias due to dependent censoring in design-based Kaplan-Meier and Cox proportional hazard model estimators. The study discusses implications of the results for survey organisations collecting event history data, researchers using surveys for event history analysis, and researchers who develop methods to correct for non-sampling biases in event history data.

Building Next Generation Consortium Services – part 3: The National Metadata Repository, Discovery Service Finna, and the New Library System

Kristiina Hormia-Poutasen esitys CBUC-konferenssissa Barcelonassa 12.4.2013.

Building Next Generation Consortium Services – part 1: Background for Change

Kristiina Hormia-Poutasen esitys CBUC-konferenssissa Barcelonassa 12.4.2013.

Building Next Generation Consortium Services – part 2: Next Generation IT-services – Why, What, When?

Kristiina Hormia-Poutasen esitys CBUC-konferenssissa Barcelonassa 12.4.2013.

Building Next Generation Consortium Services – Introduction to the seminar

Kristiina Hormia-Poutasen esitys CBUC-konferenssissa 12.4.2013 Barcelonassa.

Creating generation Y sources of motivation for enhanced performance

New challenges have been created in the modern work environment as the diversity of the workforce is greater than ever in terms of generations. There will become a large demand of generation Y employees as the baby boomer generation employees retire at an accelerated rate. The purpose of this study is to investigate Y generation specific characteristics and to identify motivational systems to enhance performance. The research questions are: 1. What are Y generation characteristics? 2. What motivational systems organizations can form to motivate Y generation employees and in turn, create better performance? The Y generation specific characteristics identified from the literature include; achievement oriented; confident; educated; multitasking; having a need for feedback; needing management support; sociable and tech savvy. The proposed motivational systems can be found in four areas of the organization; HRM, training and development, communication and decision making policies. Three focus groups were held to investigate what would motivate generation Y employees to achieve better performance. Two of these focus groups were Finnish natives and the third consisted of international students. The HRM systems included flexibility and a culture of fun. It was concluded that flexibility within the workplace and role was a great source of motivation. Culture of fun was not responded to as favorably although most focus group participants rated enjoyableness as one of their top motivating factors. Training and development systems include training programs and mentoring as sources of potential motivation. Training programs were viewed as a mode to gain a better position and were not necessarily seen as motivational systems. Mentoring programs were not concluded to have a significant effect on motivation. Communication systems included keeping up with technology, clarity and goals as well as feedback. Keeping up with technology was seen as an ineffective tool to motivate. Clarity and goal setting was seen as very important to be able to perform but not necessarily motivating. Feedback had a highly motivating effect on these focus groups. Decision making policies included collaboration and teamwork as well as ownership. Teams were familiar and meet the social needs of Y generation employees and are motivating. Ownership was equated with trust and responsibility and was highly valued as well as motivating to these focus group participants.

Combining Good Practices : Method to study Introductory Physics in Engineering Education

Julkaisumaa: 056 BE BEL Belgia

Distributed document migration between ERP systems by means of messaging and event sourcing

This bachelor’s thesis, written for Lappeenranta University of Technology and implemented in a medium-sized enterprise (SME), examines a distributed document migration system. The system was created to migrate a large number of electronic documents, along with their metadata, from one document management system to another, so as to enable a rapid switchover of an enterprise resource planning systems inside the company. The paper examines, through theoretical analysis, messaging as a possible enabler of distributing applications and how it naturally fits an event based model, whereby system transitions and states are expressed through recorded behaviours. This is put into practice by analysing the implemented migration systems and how the core components, MassTransit, RabbitMQ and MongoDB, were orchestrated together to realize such a system. As a result, the paper presents an architecture for a scalable and distributed system that could migrate hundreds of thousands of documents over weekend, serving its goals in enabling a rapid system switchover.

Luottamuksen lähteet. Vertaileva tutkimus yleistynyttä luottamusta synnyttävistä mekanismeista

Tutkimuksen aiheena on yleistynyt luottamus. Väitöskirjassa tutkitaan mistä tuntemattomien kansalaisten toisiinsa kohdistama luottamus kumpuaa ja haetaan vastauksia tähän kysymykseen sekä maakohtaisen että vertailevan tutkimuksen avulla. Tutkimus koostuu yhteenvedon lisäksi viidestä tutkimusartikkelista, joissa luottamuksen syntyä tarkastellaan sekä yksilöiden mikrotason vuorovaikutuksen että maiden välisten eroavaisuuksien näkökulmasta. Yleistyneen luottamuksen synnystä on esitetty useita eri teorioita. Tässä tutkimuksessa tarkastellaan näistä kahta keskeisintä. Osa tutkijoista korostaa kansalaisyhteiskunnan ja ruohonjuuritason verkostojen roolia yleistyneen luottamuksen synnyn taustalla. Tämän hypoteesin mukaan kansalaiset, jotka viettävät aikaansa yhdistyksissä tai muissa sosiaalisissa verkostoissa, oppivat muita helpommin luottamaan paitsi täysin tuntemattomiin ihmisiin myös yhteiskunnallisiin instituutioihin (kansalaisyhteiskuntakeskeinen hypoteesi). Toiset taas painottavat yhteiskunnan julkisten instituutioiden merkitystä. Tämä hypoteesi korostaa instituutioiden reiluutta ja oikeudenmukaisuutta (instituutiokeskeinen hypoteesi). Ihmiset pystyvät luottamaan toisiinsa ja ratkaisemaan kollektiivisia ongelmiaan yhdessä silloin kun esimerkiksi poliittiset ja lainsäädännölliset instituutiot pystyvät luomaan tähän tarvittavan toimintaympäristön. Aineistoina käytetään kansallisia (Hyvinvointi- ja palvelut) sekä kansainvälisiä vertailevia kyselytutkimuksia (European Social Survey ja ISSP). Yksilö- ja makrotason analyyseja yhdistämällä selvitetään yleistynyttä luottamusta selittäviä tekijöitä sekä mekanismeja joiden kautta yleistynyt luottamus muodostuu. Väitöskirjan tulokset tukevat suurimmaksi osaksi instituutiokeskeiseen suuntaukseen sisältyviä hypoteeseja yleistyneen luottamuksen kasautumisesta. Kuitenkin myös esimerkiksi yhdistystoiminnalla havaittiin olevan joitakin yhdistysjäsenien ulkopuolelle ulottuvia myönteisiä vaikutuksia kansalaisten luottamukseen, mikä taas tukee kansalaisyhteiskuntakeskeistä hypoteesia. Tutkimuksen keskeinen tulos on, että kaiken kaikkiaan luottamus näyttäisi kukoistavan maissa, joissa kansalaiset kokevat julkiset instituutiot oikeudenmukaisina sekä reiluina, kansalaisyhteiskunnan roolin luottamuksen synnyttämisessä ollessa tälle alisteinen. Syyksi tähän on oletettu, että näissä maissa (erityisesti pohjoismaiset hyvinvointivaltiot) harjoitettu universaali hyvinvointipolitiikka ja palvelut ovat keskeisiä korkeaa yleistynyttä luottamusta selittäviä tekijöitä. Toisaalta maavertailuissa tätä yhteyttä on selitetty myös sillä, että näissä yhteiskunnassa ei ole paikannettavissa selkeää kulttuurisesti erottuvaa alaluokkaa. Tämän tutkimuksen tulokset tukevat enemmän universaalin hyvinvointivaltion oikeudenmukaisuuteen liittyviä ominaisuuksia alaluokkaistumishypoteesin sijaan. Toisaalta mikrotasolla tarkasteltuna yleistyneen luottamuksen ja hyvinvointipalvelujen välinen yhteys liittyy enemmän palveluiden riittävyyteen kuin niiden universaalisuuden asteeseen. Niin ikään maavertailuissa esimerkiksi verotuksen oikeudenmukaisena kokeminen näyttäisi olevan palvelujen saatavuutta tai niihin liittyviä oikeudenmukaisuuden kokemuksia tärkeämpi seikka yleistyneen luottamuksen kannalta.

Reaching influential early adopters effectively in Finland : Rock Science – new generation board game

Innovations diffuse at different speed among the members of a social system through various communication channels. The group of early adopters can be seen as the most influential reference group for majority of people to base their innovation adoption decisions on. Thus, the early adopters can often accelerate the diffusion of innovations. The purpose of this research is to discover means of diffusion for an innovative product in Finnish market through the influential early adopters in respect to the characteristics of the case product. The purpose of the research can be achieved through the following sub objectives:  Who are the potential early adopters for the case product and why?  How the potential early adopters of the case product should be communicated with?  What would be the expectations, preferences, and experiences of the early adopters of the case product? The case product examined in this research is a new board game called Rock Science which is considered to be incremental innovation bringing board gaming and hard rock music together in a new way. The research was conducted in two different parts using both qualitative and quantitative research methods. This mixed method research began with expert interviews of six music industry experts. The information gathered from the interviews enabled researcher to compose the questionnaire for the quantitative part of the study. Internet survey that was sent out resulted with a sample of 97 responses from the targeted population. The key findings of the study suggest that (1) the potential early adopters for the case product are more likely to be young adults from the capital city area with great interest in rock music, (2) the early adopters can be reached effectively through credible online sources of information, and (3) the respondents overall product feedback is highly positive, except in the case of quality-price ratio of the product. This research indicates that more effective diffusion of Rock Science board game in Finland can be reached through (1) strategic alliances with music industry and media partnerships, (2) pricing adjustments, (3) use of supporting game formats, and (4) innovative use of various social media channels.

INSPIRE: contributions to Invenio from the leading High Energy Physics (HEP) platform

Presentation at Open Repositories 2014, Helsinki, Finland, June 9-13, 2014

Biomedical Event Extraction with Machine Learning

Biomedical natural language processing (BioNLP) is a subfield of natural language processing, an area of computational linguistics concerned with developing programs that work with natural language: written texts and speech. Biomedical relation extraction concerns the detection of semantic relations such as protein-protein interactions (PPI) from scientific texts. The aim is to enhance information retrieval by detecting relations between concepts, not just individual concepts as with a keyword search. In recent years, events have been proposed as a more detailed alternative for simple pairwise PPI relations. Events provide a systematic, structural representation for annotating the content of natural language texts. Events are characterized by annotated trigger words, directed and typed arguments and the ability to nest other events. For example, the sentence “Protein A causes protein B to bind protein C” can be annotated with the nested event structure CAUSE(A, BIND(B, C)). Converted to such formal representations, the information of natural language texts can be used by computational applications. Biomedical event annotations were introduced by the BioInfer and GENIA corpora, and event extraction was popularized by the BioNLP'09 Shared Task on Event Extraction. In this thesis we present a method for automated event extraction, implemented as the Turku Event Extraction System (TEES). A unified graph format is defined for representing event annotations and the problem of extracting complex event structures is decomposed into a number of independent classification tasks. These classification tasks are solved using SVM and RLS classifiers, utilizing rich feature representations built from full dependency parsing. Building on earlier work on pairwise relation extraction and using a generalized graph representation, the resulting TEES system is capable of detecting binary relations as well as complex event structures. We show that this event extraction system has good performance, reaching the first place in the BioNLP'09 Shared Task on Event Extraction. Subsequently, TEES has achieved several first ranks in the BioNLP'11 and BioNLP'13 Shared Tasks, as well as shown competitive performance in the binary relation Drug-Drug Interaction Extraction 2011 and 2013 shared tasks. The Turku Event Extraction System is published as a freely available open-source project, documenting the research in detail as well as making the method available for practical applications. In particular, in this thesis we describe the application of the event extraction method to PubMed-scale text mining, showing how the developed approach not only shows good performance, but is generalizable and applicable to large-scale real-world text mining projects. Finally, we discuss related literature, summarize the contributions of the work and present some thoughts on future directions for biomedical event extraction. This thesis includes and builds on six original research publications. The first of these introduces the analysis of dependency parses that leads to development of TEES. The entries in the three BioNLP Shared Tasks, as well as in the DDIExtraction 2011 task are covered in four publications, and the sixth one demonstrates the application of the system to PubMed-scale text mining.

On advancing business intelligence in the electricity retail market

In recent decades, business intelligence (BI) has gained momentum in real-world practice. At the same time, business intelligence has evolved as an important research subject of Information Systems (IS) within the decision support domain. Today’s growing competitive pressure in business has led to increased needs for real-time analytics, i.e., so called real-time BI or operational BI. This is especially true with respect to the electricity production, transmission, distribution, and retail business since the law of physics determines that electricity as a commodity is nearly impossible to be stored economically, and therefore demand-supply needs to be constantly in balance. The current power sector is subject to complex changes, innovation opportunities, and technical and regulatory constraints. These range from low carbon transition, renewable energy sources (RES) development, market design to new technologies (e.g., smart metering, smart grids, electric vehicles, etc.), and new independent power producers (e.g., commercial buildings or households with rooftop solar panel installments, a.k.a. Distributed Generation). Among them, the ongoing deployment of Advanced Metering Infrastructure (AMI) has profound impacts on the electricity retail market. From the view point of BI research, the AMI is enabling real-time or near real-time analytics in the electricity retail business. Following Design Science Research (DSR) paradigm in the IS field, this research presents four aspects of BI for efficient pricing in a competitive electricity retail market: (i) visual data-mining based descriptive analytics, namely electricity consumption profiling, for pricing decision-making support; (ii) real-time BI enterprise architecture for enhancing management’s capacity on real-time decision-making; (iii) prescriptive analytics through agent-based modeling for price-responsive demand simulation; (iv) visual data-mining application for electricity distribution benchmarking. Even though this study is from the perspective of the European electricity industry, particularly focused on Finland and Estonia, the BI approaches investigated can: (i) provide managerial implications to support the utility’s pricing decision-making; (ii) add empirical knowledge to the landscape of BI research; (iii) be transferred to a wide body of practice in the power sector and BI research community.