University of Twente Student Theses
Adapted CRISP-DM approach for recommendation system development for most suitable open-source ETL tool
Grotentraast, Jurgen (2024) Adapted CRISP-DM approach for recommendation system development for most suitable open-source ETL tool.
This is the latest version of this item.
PDF
1MB |
Abstract: | This paper covers the adaptation of the CRoss Industry Standard Process for Data Mining (CRISP-DM) methodology for developing a recommendation system for open-source ETL tools. This recommendation system helps users find a suitable ETL tool for their use case. Therefore, interviews were conducted with developers to find the key aspects of ETL tools among others security, ETL pipeline design, and hosting. Different open-source tools were analyzed on these key aspects in the form of an aspect matrix. This aspect matrix was transformed into a scorecard to filter and rank different tools based on requirements. A questionnaire was created to gather the requirements and provide recommendations to the user. Lastly, the recommendation system was evaluated in three ways. As a form of self-reflection compliance with the seven guidelines by Hevner et al. was rated to reflect on the development process. The recommendation system was evaluated with a survey in which participants could use it to rate its understandability, usability, and clarity. Overall, participants rated the recommendation system a 6.8 out of 10. The main improvements could be made in the presentation and motivation of the recommendations. The results indicate the adaptations made to the CRISP-DM methodology were appropriate and useable for developing a recommendation system. |
Item Type: | Essay (Master) |
Faculty: | EEMCS: Electrical Engineering, Mathematics and Computer Science |
Subject: | 54 computer science |
Programme: | Computer Science MSc (60300) |
Link to this item: | https://purl.utwente.nl/essays/104732 |
Export this item as: | BibTeX EndNote HTML Citation Reference Manager |
Repository Staff Only: item control page