1,465 research outputs found
An in-depth investigation on the behavior of measures to quantify reproducibility
No full textScience is facing a so-called reproducibility crisis, where researchers struggle to repeat experiments and to get the same or comparable results. This represents a fundamental problem in any scientific discipline because reproducibility lies at the very basis of the scientific method. A central methodological question is how to measure reproducibility and interpret different measures. In Information Retrieval (IR), current practices to measure reproducibility rely mainly on comparing averaged scores. If the reproduced score is close enough to the original one, the reproducibility experiment is deemed successful, although the identical scores can still rely on entirely different result lists. Therefore, this paper focuses on measures to quantify reproducibility in IR and their behavior. We present a critical analysis of IR reproducibility measures by synthetically generating runs in a controlled experimental setting, which allows us to control the amount of reproducibility error. These synthetic runs are generated by a deterioration algorithm based on swaps and replacements of documents in ranked lists. We investigate the behavior of different reproducibility measures with these synthetic runs in three different scenarios. Moreover, we propose a normalized version of Root Mean Square Error (RMSE) to quantify reproducibility better. Experimental results show that a single score is not enough to decide whether an experiment is successfully reproduced because such a score depends on the type of effectiveness measure and the performance of the original run. This study highlights how challenging it can be to reproduce experimental results and quantify the amount of reproducibility.</p
Formalised Information Needs dataset
No full textDataset for the following publication:
@inproceedings{JCDL23_KreutzBSSW,
author = {Christin Katharina Kreutz and
Martin Blum and
Philipp Schaer and
Ralf Schenkel and
Benjamin Weyers},
title = {Evaluating Digital Library Search Systems by using Formal Process Modelling},
booktitle = {{JCDL} '23}
}
The dataset contains evaluation data from an user study with 13 participants working on two tasks: expert search (T_ex) and paper search (T_pa). For participants there are three BPMN models per task, depicting 1) their ideal task conduction model (vIMM), 2) the expert-generated translation of this model to the SchenQL digital library system (vPGM) and 3) the participants' task conduction model using SchenQL (PCM). These BPMNs are given as SVGs in folders corresponding to the tasks(T_ex and T_pa) inside folders corresponding to participants' code names.
For all participants the transcript of the first user session is contained as text files in a folder named Transcripts Session 1. Additionally, participants' responses to the questionnaires are included (in Questionnaires.csv)
Philipp Melanchthon
Get PDFPhilipp Melanchthon (1497–1560) was, with Martin Luther, the most influential reformer of the church during the 16th century. He was also a reformer of university education, especially theological studies, as well as the school system in Germany. He was responsible for a theological curriculum that included Greek, Hebrew, and philosophy. He, as a professor of Greek at the University of Wittenberg since 1518, was the author of the first generally accepted Protestant confession, known as the Confessio Augustana (1530). He also wrote the first Protestant commentaries on Paul’s letter to the Romans (1519), as well as the first Protestant handbook in systematic theology (1521). He was the main negotiator of the Protestant movement during the diets and religious discussions with the Roman Catholic Church. He is known as the ‘teacher of Germany and Europe’ and is respected as the father of the ecumenical movement. Yet, Melanchthon is not known to South Africans and especially Afrikaans-speaking people who, traditionally, have close links with the reformational tradition. There is not yet one single publication on Melanchthon in Afrikaans or by a South African scholar, making this book, therefore, the first by an Afrikaans-speaking scholar on Melanchthon
#Bitcoin : an analysis of the field of a decentralized virtual currency using twitter data
No full textAuthor Philipp AllerstorferAbstract in englischer SpracheMasterarbeit Universität Linz 201
On Philipp Lersch’s Psychology, Reflection No. 2
No full textАвтор статті зазначає, що у 1930–1970 рр. Ф. Лерш був головним представником гуманітарної течії в німецькомовній психології. Наукову класифікацію людських прагнень і стремлінь було
викладено у головній праці Ф. Лерша «Структура особи» (11-те видання 1970 року). Ф. Лерш зумів
об’єднати у своїм творі інформацію із низки наук (біології, психопатології, із різних психологічних
теорій та концепцій).The author of the article notes that in 1930-1970, Philipp Lersch was the main representative of the
humanitarian direction in German psychology. A scientific classification of human aspirations and ambitions
was described in the main work by Philipp Lersch “Building the Person” (11th ed. of 1970). In his work,
Philipp Lersch managed to combine information from a number of sciences (biology, psychopathology, various
psychological theories and concepts)
Deformation Behavior and Damage Modeling of Polypropylene and Polycarbonate
No full textAuthor Philipp Siegfried StelzerKurzfassungen in deutscher und englischer SpracheMasterarbeit Universität Linz 201
Philipp Timischl x Numeró Art
No full textThis article discusses the intersections of gender and class drag in the work of Austrian artist Philipp Timischl. The essay takes the form of a letter of refusal, discussing why the author is simply too busy, then goes on to discuss other performances of industriousness, from Kevin Floyd's proposal that modern masculinity is a labour based performance through to the artist's own gender enhancing or swapping works
Informetric Analyses and Non-textual Document Attributes for Information Retrieval in Digital Libraries
No full textDie Suche nach wissenschaftlicher Literatur ist eine Forschungsherausforderung für das Information Retrieval im besonderen Umfeld der digitalen Bibliotheken. Aktuelle Nutzerstudien zeigen, dass im klassischen IR-Modell zwei typische Schwächen auszumachen sind: das Ranking der gefundenen Dokumente und Probleme bei der Formulierung von Suchanfragen. Gleichzeitig ist zu sehen, dass traditionelle Retrievalsysteme, die primär textuelle Dokument- und Anfragemerkmale nutzen, bei IR-Evaluationskampagnen wie TREC und CLEF in ihrer Leistung seit Jahren stagnieren.
Zwei informetrisch-motivierte Verfahren zur Suchunterstützung werden vorgestellt und mittels einer Laborevaluation mit den beiden IR-Testkollektionen GIRT und iSearch sowie 150 und 65 Topics evaluiert. Die Verfahren sind: (1) eine auf der Kookkurrenz von Dokumentattributen basierende Anfrageerweiterung und (2) ein Rankingansatz, der informetrische Beobachtungen zur Produktivität von Informationserzeugern ausnutzt. Beide Verfahren wurden mit einer Referenzimplementation auf Basis der Suchmaschine Solr verglichen. Beide Verfahren zeigen positive Effekte beim Einsatz von zusätzlichen Dokumentattributen wie Autorennamen, ISSN-Codes und kontrollierten Schlagwörtern. Bei der Anfrageerweiterung konnte ein positiver Effekt in Form einer Verbesserung der Precision (bpref +12%) und des Recall (R +22%) erzielt werden. Die alternativen Rankingansätze konnten beim Ansatz von Autorennamen und ISSN-Codes die Baseline erreichen bzw. diese beim Einsatz der kontrollierten Schlagwörter über- treffen (MAP +14%). Einen negativen Einfluss auf das Ranking hatten allerdings die Einbeziehung von Faktoren wie Verlagsnamen oder Erscheinungsorten. Für beide Verfahren konnte eine substantiell andere Sortierung der Ergebnismenge, gemessen anhand von Kendalls, beobachtet werden. Zusätzlich zu der verbesserten Relevanz der Ergebnisliste kann der Nutzer so eine neue Sicht auf die Dokumentenmenge gewinnen.
Die Anfrageerweiterung mit Autorennamen, ISSN-Codes und Thesaurustermen zeigt das bisher ungenutzte Potential, das sich in digitalen Bibliotheken durch die Datenfülle und -qualität ergibt. Die Rankingverfahren konnten die Leistung des Baseline-Systems übertreffen, nachdem eine Überprüfung auf Vorliegen einer Power Law-Verteilung und eine anschließende Filterung durchgeführt wurde. Dies zeigt, dass die Rankingverfahren nicht universell für alle Suchanfragen anwendbar sind, sondern ein Vorhandensein bestimmter Häufigkeitsverteilungen voraussetzen. So wird die enge Verbindung der Verfahren zu informetrischen Gesetzmäßigkeiten wie Bradfords, Lotkas oder Zipfs Gesetz deutlich. Die beiden in der Arbeit evaluierten Verfahren sind als interaktive Suchunterstützungsdienste in der sozialwissenschaftlichen digitalen Bibliothek Sowiport implementiert. Die Verfahren lassen sich über entsprechende Web- Schnittstellen auch in anderen Anwendungskontexten einsetzen.The search for scientific literature in scientific information systems is a discipline at the intersection between information retrieval and digital libraries. Recent user studies show two typical weaknesses of the classical IR model: ranking of retrieved and maybe relevant documents and the language problem during the query formulation phase. At the same time traditional retrieval systems that rely primarily on textual document and query features are stagnating for years, as it could be observed in IR evaluation campaigns such as TREC or CLEF. Therefore alternative approaches to surpass these two problem fields are needed. Two different search support systems are presented in this work and evaluated with a lab evaluation using the IR test collection GIRT and iSearch with 150 and 65 topics, respectively. These two systems are (1) a query expansion that is based on the analysis of co-occurrences of document attributes and (2) a ranking mechanism that applies informetric analysis of the productivity of information producers in the information production process. Both systems were compared to a baseline system using the Solr search engine. Both methods showed positive effects when applying additional document attributes like author names, ISSN codes and controlled terms. The query expansion showed an improvement in precision (bpref +12%) and in recall (R +22%).
he alternative ranking methods were able to compete with the baseline for author names and ISSN codes and were able to beat the baseline by using controlled terms (MAP +14%). A clear negative influence was seen when using entities like publishers or locations. Both methods were able to generate a substantially different sorting of the result set, measured using Kendall. So, additional to the improved relevance in the result list, the user can get a new and different view on the document set. Query expansion using author names, ISSN codes and thesaurus terms showed great potential that lies within the rich metadata sets of digital library systems. The proposed ranking methods could outperform standard relevance ranking methods after they were filtered by the existence of a so-called power law. This showed that the proposed ranking methods cannot be used universally in any case but require specific frequency distributions in the metadata. A connection between the underlying informetric laws of Bradford, Lotka and Zipf is made clear. The evaluated methods were implemented as interactive search supporting systems that can be used in an interactive prototype and the social science digital library system Sowiport. Besides that, the methods are adaptable to other systems and environments using a free software framework and a web API
Author Correction: Auto-aggressive CXCR6+ CD8 T cells cause liver immune pathology in NASH
No full textIn this Article, the surname of Tobias Boettler was incorrectly shown as ‘Böttler’, and the surname of author Jan-Philipp Mallm was incorrectly shown as ‘Malm’. The original Article has been corrected online
- …
