For the future demand prediction of identification documents the National Office for Identity Data is interested in a new prediction model based on machine learning techniques. Due to the existence of many different machine learning algorithms and the often competing model requirements regarding model performance and explainability it can be difficult to determine which machine learning algorithm to use. This research aims to find a solution to help decision-makers determine what machine learning algorithms they should use while having to account for competing model requirements. This is done by performing a case study for the National Office for Identity Data in which the use of machine learning for a new prototype prediction model is explored. Additionally, a novel explainability assessment table is designed through which the explored machine learning algorithms are assessed on their explainability. By combining the results from the case study and the explainability assessment analysis a final design is created in the form of a decision support framework. This framework aims to support decision-makers in their process of finding a suitable machine learning algorithm when dealing with competing model requirements. Future researchers are recommended to either extend upon the decision support framework by making it more accessible for people with a less technical background or to analyse how a more standardised definition of model explainability can be developed and implemented in the field of machine learning.

Healthcare recommender systems emerged to help patients make better decisions for their health, leveraging the vast amount of data and patient experience. One type of this system focuses on recommending the most appropriate physician based on previous patient feedback in the form of ratings. Such advice can be challenging to generate for new patients as their interaction with the new environment is too limited to inform meaningful recommendations. Their preferences from the previous system can be transferred to the new system to help generate recommendations in a cross-domain approach.
Cross-domain recommender systems aim to bridge the gap between domains where such interactions are plentiful and domains with relatively little interactions for the new users, using transfer learning techniques. However, while privacy concerns in single-domain recommender systems have been studied extensively, the same does not apply to cross-domain systems where multiple domains collaborate. Privacy is essential in settings such as healthcare, where rating information is sensitive and prone to leaking patient medical history to third parties. Existing works lack an efficient design that provides sufficient privacy while generating recommendations.
This thesis presents a privacy-preserving cross-domain recommender system with partial user overlap. We first propose a privacy-preserving matrix factorization approach with differential privacy and multiparty setting to serve as the base of our cross-domain design. Our matrix factorization design with privacy is easy to adopt by existing systems, incurring negligible overhead, and requires no modification to existing factorization implementations. Next, we show the privacy-accuracy trade-off in our base method with experiments over the MovieLens dataset. Finally, in the cross-domain setup, we demonstrate that adding privacy to the proposed scheme has a minimal predictive performance loss, essentially providing privacy for free.

The goal of this research is to improve the fitness for use of the CMDB in the development phase of large infrastructure projects by identifying the information needs of the individuals. The construction industry is transiting from paper-based systems engineering to Model-Based Systems Engineering, especially in the development phase of large infrastructure projects. The configuration management database, which manages the systems engineering information, is considered to be a the central systems in the Model Based Systems Engineering development. In current practice not all individuals perceive the benefits of the implementation of systems engineering tools. The context study shows the fundamental differences in systems engineering processes consisting of requirements analysis, functional analysis, design synthesis, verification and validation, system analysis and control. Based on the Integral Project Model, the technical management department has the biggest stake in the development phase. The roles considered in the technical management department are the specialty engineer and the lead engineer. The literature study presents the information needs characteristics (individual demographics, context, frequency, predictability, importance and complexity) as well as the fitness for use assessment method. The practical study shows that the information needs of the particular users differ per process as the roles have different tasks and responsibilities. As a result, the fitness for use of the configuration management database is also considered differently. The specialty engineer mainly requires information for decision making whereas the lead engineer focusses on management activities. The specialty engineer, which is mainly involved in the design synthesis and the system analysis processes, experiences problems obtainability, completeness, timeliness and coherence of the information. As a result, the specialty engineer implements workarounds, which might cause a snowball effect. The lead engineer mainly experiences problems with the functionalities of the system in the requirements and functional analysis. As a result, the lead engineer implements workarounds such as Word or Excel. As the specialty engineer experiences the majority of the problems, the design of solutions focusses on the specialty engineer. The solutions proposed consists of establishing physical relations between objects, implementation of abstraction levels, information presentation in the shape of a one stop page and information presentation in the shape of custom queries. The physical relations between objects and the abstract levels enable the one stop page and custom queries. The one stop query meets the predictable and recurrent information needs of the specialty engineer, whereas the custom query meets the unpredictable, new, urgent and complex information needs. The practical relevance of the solutions has been explained by a use case example in the shape of a simplified tunnel project. The relevance of the one stop page and custom query solutions are illustrated by stating the information needs in the particular processes and how the solutions meet these needs.

After the start of the COVID19 pandemic, many countries decided to develop digital contact tracing applications to assist in contact tracing. This aimed to warn people who were potentially infected as early as possible and stop them from spreading the virus further and bringing down the total amount of new infections. Even though there was some proof of the potential effectiveness of applications such as these, they had never been deployed at scale before the pandemic and there was no standardised and shared vision on how to best implement these applications, meaning that countries that wanted to include digital contact tracing as part of their COVID19 strategies needed to design and develop an application themselves...

In recent years, false information on social media platforms has become a centre of attention due to its consequences on elections and public health. COVID-19 has shown how an infodemic fuelled by false information can be detrimental to public health. Existing measures are ineffective. Thus, the thesis takes an exploratory approach from the lens of new institutional economics. A comparative institutional analysis between public health institutions and social media platforms concerning their information discovery process and how differences between them can lead to false information on social media platforms is performed. The findings question whether false information is as widespread on social media platforms as projected, identifies the role of public health institutions and politicians in the spread of false information, how information is regulated on social media platforms and what drives the information discovery process on social media platforms. Based on the comparative institutional analysis, it is recommended to implement prediction markets to address false information on social media platforms.

When discussing media, parasocial phenomena is mentioned. Audience members naturally partake in parasocial phenomena when consuming media and empathizing with media figures. Viewers who form relationships with a media figure despite never meeting them are said to be in a parasocial relationship. On YouTube, examples of parasocial interactions occur when content creators request likes, subscriptions, or thank viewers for watching.
While existing literature recognizes that viewers form parasocial relationships with YouTube content creators, there is not enough emphasis on understanding the content creators’ perspectives.
This thesis helps to understand how content creators experience being the target of parasocial relationships. Furthermore, the general feelings and emotions of content creators are probed in order to extract the values of content creators, such that content creators’ unique experiences can be analyzed and educated recommendations can be given to relevant stakeholders.
Eight interviews were held with content creators who are the face of their channel. Three main research questions were formulated: (RQ2) How and to what extent do content creators experience parasocial phenomena on the YouTube platform? (RQ3) What moral challenges do content creators experience when connecting with their audience or expressing themselves in their content? (RQ4) How does the YouTube platform influence the ways in which content creators are able to communicate their values to their audience?
To ensure these research questions were answered, a set of propositional statements were formulated from the existing literature. Each proposition consists of one specific element which may be supported by interview findings. When compounded, these propositions answer their relevant research question.
For RQ2, propositions include Interviewee holds influence over their audience, has met a viewer in person, has audience members similar to themselves, and does not have the same amount of knowledge about the viewer as the viewer has about them. For RQ3, propositions include Interviewee interprets their audience as a group rather than as individuals, communicates with their audience members outside of their video content, feels they need to be professional in their YouTube content, creates videos as a creative outlet, creates videos because it is fun for them, and feels connected to their audience. For RQ4, propositions include Interviewee is aware of the YouTube algorithm affecting the platform, receives financial benefits or a stream of income from YouTube, feels authentic on YouTube, is exposed to tools to grow their YouTube audience, and has confronted their values when making decisions on YouTube.
Interviewees’ anecdotal evidence directly support findings that YouTube content creators value at least one or more of the following: self-expression, creativity, loyalty, community, credibility, authenticity, entertainingness, education, fiscal opportunity, social justice, content quality, pride in self or work, appreciation for viewers, and filling a niche.
YouTube is a socio-technical system with an immeasurable number of users every single day. YouTube should be honest with their content creators and viewers, as the YouTube platform would not exist in its current state without them.
Calls for future research and explicit recommendations are given to relevant stakeholders about how to handle parasocial phenomena.