Schnitzer, Steffen (2019)
Task Recommendation in Crowdsourcing Platforms.
Technische Universität Darmstadt
Dissertation, Erstveröffentlichung
Kurzbeschreibung (Abstract)
Task distribution platforms, such as micro-task markets, project assignment portals, and job search engines, support the assignment of tasks to workers. Public crowdsourcing platforms support the assignment of tasks in micro-task markets to help task requesters to complete their tasks and allow workers to earn money. Enterprise crowdsourcing platforms provide a marketplace within enterprises for the internal placement of tasks from employers to employees. Most of both types of task distribution platforms rely on the workers' selection capabilities or provide simple filtering steps to reduce the number of tasks a worker can choose from. This self-selection mechanism unfortunately allows for tasks to be performed by under- or over-qualified workers. Supporting the workers by introducing a task recommender system helps to solve such deficits of existing task distributions.
In this thesis, the requirements towards task recommendation in task distribution platforms are gathered with a focus on the worker's perspective, the design of appropriate assignment strategies is described, and innovative methods to recommend tasks based on their textual descriptions are provided. Different viewpoints are taken into account by analyzing the domains of micro-tasks, project assignments, and job postings. The requirements of enterprise crowdsourcing platforms are compiled based on the literature and a qualitative study, providing a conceptual design of task assignment strategies. The demands of workers and their perception of task similarity on public crowdsourcing platforms are identified, leading to the design and implementation of additional methods to determine the similarity of micro-tasks. The textual descriptions of micro-tasks, projects, and job postings are analyzed in order to provide innovative methods for task recommendation in these domains.
Typ des Eintrags: | Dissertation | ||||
---|---|---|---|---|---|
Erschienen: | 2019 | ||||
Autor(en): | Schnitzer, Steffen | ||||
Art des Eintrags: | Erstveröffentlichung | ||||
Titel: | Task Recommendation in Crowdsourcing Platforms | ||||
Sprache: | Englisch | ||||
Referenten: | Steinmetz, Prof. Dr. Ralf ; Rensing, PD Dr. Christoph | ||||
Publikationsjahr: | 2019 | ||||
Ort: | Darmstadt | ||||
Datum der mündlichen Prüfung: | 12 Februar 2019 | ||||
URL / URN: | https://tuprints.ulb.tu-darmstadt.de/8549 | ||||
Kurzbeschreibung (Abstract): | Task distribution platforms, such as micro-task markets, project assignment portals, and job search engines, support the assignment of tasks to workers. Public crowdsourcing platforms support the assignment of tasks in micro-task markets to help task requesters to complete their tasks and allow workers to earn money. Enterprise crowdsourcing platforms provide a marketplace within enterprises for the internal placement of tasks from employers to employees. Most of both types of task distribution platforms rely on the workers' selection capabilities or provide simple filtering steps to reduce the number of tasks a worker can choose from. This self-selection mechanism unfortunately allows for tasks to be performed by under- or over-qualified workers. Supporting the workers by introducing a task recommender system helps to solve such deficits of existing task distributions. In this thesis, the requirements towards task recommendation in task distribution platforms are gathered with a focus on the worker's perspective, the design of appropriate assignment strategies is described, and innovative methods to recommend tasks based on their textual descriptions are provided. Different viewpoints are taken into account by analyzing the domains of micro-tasks, project assignments, and job postings. The requirements of enterprise crowdsourcing platforms are compiled based on the literature and a qualitative study, providing a conceptual design of task assignment strategies. The demands of workers and their perception of task similarity on public crowdsourcing platforms are identified, leading to the design and implementation of additional methods to determine the similarity of micro-tasks. The textual descriptions of micro-tasks, projects, and job postings are analyzed in order to provide innovative methods for task recommendation in these domains. |
||||
Alternatives oder übersetztes Abstract: |
|
||||
URN: | urn:nbn:de:tuda-tuprints-85496 | ||||
Sachgruppe der Dewey Dezimalklassifikatin (DDC): | 000 Allgemeines, Informatik, Informationswissenschaft > 004 Informatik | ||||
Fachbereich(e)/-gebiet(e): | 20 Fachbereich Informatik | ||||
Hinterlegungsdatum: | 14 Apr 2019 19:55 | ||||
Letzte Änderung: | 14 Apr 2019 19:55 | ||||
PPN: | |||||
Referenten: | Steinmetz, Prof. Dr. Ralf ; Rensing, PD Dr. Christoph | ||||
Datum der mündlichen Prüfung / Verteidigung / mdl. Prüfung: | 12 Februar 2019 | ||||
Export: | |||||
Suche nach Titel in: | TUfind oder in Google |
Frage zum Eintrag |
Optionen (nur für Redakteure)
Redaktionelle Details anzeigen |