Fair team recommendations for multidisciplinary projects
Machado, Lucas (2019)
Machado, Lucas
2019
Tietojenkäsittelytieteiden tutkinto-ohjelma - Degree Programme in Computer Sciences
Informaatioteknologian ja viestinnän tiedekunta - Faculty of Information Technology and Communication Sciences
This publication is copyrighted. You may download, display and print it for Your own personal use. Commercial use is prohibited.
Hyväksymispäivämäärä
2019-05-13
Julkaisun pysyvä osoite on
https://urn.fi/URN:NBN:fi:tuni-201905161729
https://urn.fi/URN:NBN:fi:tuni-201905161729
Tiivistelmä
With the ever increasing amount of data in the world, it becomes harder to find useful and desired information. Recommender systems, which offer a way to analyze that data and suggest relevant information, are already common nowadays and a important part of several systems and services. While recommender systems are often used for suggesting items for users, there are not many studies about using them for problems such as team formation. This thesis focus on exploring a variation of that problem, in which teams have multidisciplinary requirements and members' selection is based on the match of their skills and the requirements. In addition, when assembling multiple teams there is a challenge of allocating the best members in a fair way between the teams.
With the studied concepts from the literature, this thesis suggests a brute force and a faster heuristic method as solutions to create team recommendations to multidisciplinary projects. Furthermore, to increase the fairness between the recommended teams, the K-rounds and Pairs-rounds methods are proposed as variations of the heuristic approach.
Several different test scenarios are executed to analyze and compare the efficiency and efficacy of these methods, and it is found that the heuristic-based methods are able to provide the same levels of quality with immensely greater performance than the brute force approach. The K-rounds method is able to generate substantially more fair team recommendations, while keeping the same levels of quality and performance as other methods. The Pairs-rounds method presents slightly better recommendations quality-wise than the K-rounds method, but its recommendations are less fair to a small degree. The proposed methods perform well enough for use in real scenarios.
With the studied concepts from the literature, this thesis suggests a brute force and a faster heuristic method as solutions to create team recommendations to multidisciplinary projects. Furthermore, to increase the fairness between the recommended teams, the K-rounds and Pairs-rounds methods are proposed as variations of the heuristic approach.
Several different test scenarios are executed to analyze and compare the efficiency and efficacy of these methods, and it is found that the heuristic-based methods are able to provide the same levels of quality with immensely greater performance than the brute force approach. The K-rounds method is able to generate substantially more fair team recommendations, while keeping the same levels of quality and performance as other methods. The Pairs-rounds method presents slightly better recommendations quality-wise than the K-rounds method, but its recommendations are less fair to a small degree. The proposed methods perform well enough for use in real scenarios.