• Learning Analytics

Liaqat Ali

Leveraging MSLQ dataset for predicting students’ achievement goal orientations

Motivation, cognition, and achievement goals are three broad domains of learners’ characteristics that affect how learners study and what they learn by studying. Two of the most commonly used instruments for measuring learners’ characteristics are the Motivated Strategies for Learning Questionnaire, and the Achievement Goals Orientation. A substantial body of research over the last three decades has studied relationships between the motivational and the achievement goal constructs used in both the instruments. No previous study, however, attempted to use the existing knowledge of construct associations to derive learners’ achievement goals from their measures of learning motivation or vice versa. This research aimed to leverage the Motivated Strategies for Learning Questionnaire dataset for predicting learners’ achievement goals orientations and was guided by the following research question: whether the MSLQ measures of motivated strategies for learning reveal achievement goal orientations of college students. Data for this study was collected from 347 undergraduate students. Both a confirmatory data analysis approach and an exploratory data analysis approach were employed to examine the collected data. For confirmatory analysis, I built a new theoretical model of the Motivated Strategies for Learning Questionnaire items based on the previous empirical research findings, and employed Pearson correlation analysis, regression analysis and Akaike Information Criterion to identify the best-fit models. For exploratory investigations, I used canonical correlation analysis to identify relationships between Motivated Strategies for Learning Questionnaire measures and Achievement Goals Orientation constructs. The confirmatory analysis identified a 15-item model of motivated strategies which predicted four achievement orientations, whereas the exploratory analysis resulted in a 15-item model that predicted three achievement goal orientations.

Full Thesis

• Learning Analytics
• Open Learner Models

Sanam Shirazi

Varying Effects of Learning Analytics Visualizations for Students with Different Achievement Goal Orientations

Description: Through advancements of Technology-Enhanced Learning an opportunity has emerged to provide students with timely feedback using Learning Analytics in the form of visualizations. To afford actual impact on learning, such tools have to be informed by theories of education. Particularly, educational research shows that individual differences play a significant role in explaining students’ learning process. However, limited empirical research has investigated the role of theoretical constructs such as motivational factors that are underlying the observed differences between individuals. In this work, we conducted a field experiment to examine the effect of three designed Learning Analytics Visualizations on students’ participation in online discussions in authentic course settings. Using hierarchical linear mixed models, our results revealed different effects of visualizations on the quantity and quality of messages posted by students with different Achievement Goal Orientations. Findings highlight the methodological importance of considering individual differences and pose important implications for future design of Learning Analytics Visualizations.

Full Thesis | LinkedIn

• Software Engineering

Mohsen Asadi

Developing and Validating Customizable Process Models

Description: A process model defines the activities of a business process and their attributes (e.g. cost and time). Process models typically are instantiated several times and every process instance may be executed differently based on the context and the requirements of target stakeholders. Hence, several variants of the same process model may coexist in organizations which urges the organizations to support flexible processes in order to cope with process variabilities. Motivated by the need of flexible process models, a number of approaches have been proposed for the development of customizable process models which integrate variability in process models.The development and adaptation of customizable process models raise several challenges: 1) the need for taking into account several variability types (i.e. OR, alternative, and optional) which may occur in a customizable process model; 2) integrating variability into process models impose additional modeling complexity; 3) deriving a process variant from a customizable process model requires the close consideration of a target application requirements and relation between variants and requirements; 4) ensuring compliance of a process variant with behavioral and configuration constraints formulated in a customizable process model.

This dissertation presents a feature oriented customization and validation framework for customizable process models. The customization component relies on software product lines and utilizes feature modeling techniques for modeling variability in customizable process models. Additionally, a pre-configuration process, a decision making technique called Stratified Analytical Hierarchy Process (S-AHP), and Artificial Intelligent Planning Techniques are provided to derive a process variant from a customizable process model based on the stakeholders requirements. The validation component identifies a set possible inconsistency patterns between requirements model (represented by goal model), variability model (represent by feature model), and customizable process models and employs Description Logic to detect the inconsistencies.We evaluated the framework using a set of experiments and explored the running time of the proposed techniques under different sizes of models and constraints. The results show that the running time of proposed techniques is tractable in practical customizable process models. Additionally, a comparative analysis of the components of the framework is conducted which reveals improvements over state of the art in customizable process models.

Full Thesis | LinkedIn

• Software Engineering

Bardia Mohabbati

Quality-aware Service-Oriented Software Product Lines: Feature-Driven Process Configuration and Optimization

Description: Research initiative in Service-Oriented Computing aims at developing adaptable and scalable distributed applications and addressing challenges such as application integration, reusability, modularity, and interoperability. Service-Oriented Architecture (SOA) as an architectural style enables organizations to offer their application functionality as a service and enhance the adaptability to changes of new requirements of stakeholders, i.e., service consumers. Nowadays enterprises and service providers face several challenges to develop SOA-based solutions. They indispensably require to effectively manage variability in both functional and non-functional (quality) requirements at the business process level to rapidly and cost-effectively develop and deploy customized services that best meet the stakeholders' feature needs. SOAs provide the architectural underpinnings to support software reuse and enable variability at both design and run-time; however, they lack support to manage variability that promotes configurability and customization. Variability modeling and management have been the core research subjects in Software Product Line Engineering (SPLE) with the objective of addressing the issues of engineering and developing software-intensive systems. Combining SPLE and SOAs has been a subject of considerable research interest in recent years to develop highly configurable software systems.

In this thesis, we adopt a product-line approach in the service domain and hypothesize that the SPLE paradigm, enabling variability management and systematic planned reuse, can be applied orthogonally to aid Service-Oriented Software Engineering (SOSE) to yield these benefits and construct Service-Oriented Software Product Lines (SOSPLs). We propose the Configurable Process Models as the realization of SOSPLs, where services are the building blocks for the implementation of software features, which provide support for variation among members of a product line configured based on stakeholders' requirements.Our proposed approach provides scalable and efficient automated decision-making support in the course of configuration helping to create tailored software services according to stakeholders’ preferences.

The key contributions of this thesis are: (i) a systematic analysis of the state-of-the-art research; (ii) a methodology to support variability modeling and management for the development of an SOSPL; (iii) a quality model and evaluation method; (iv) a framework supporting automatic quality-aware process configuration; and (v) an empirical evaluation of performance and scalability of approach.

Full Thesis | LinkedIn

• Learning Analytics

Melody Siadaty

Semantic Web-Enabled Interventions to Support Workplace Learning

Description: To keep pace with today’s rapidly growing knowledge-driven society, continuous learning in workplaces and being able to self-regulate one’s learning processes have become essential. In this dissertation, I propose a set of interventions, developed using Semantic-Web technologies, to scaffold self-regulated learning (SRL) processes in workplaces. I integrate social embeddedness elements with harmonization components in the functionalities provided by these interventions to accentuate the social and contextual dimensions of workplace learning. To measure users’ SRL processes, I developed a trace-based protocol which captures users’ low-level trace data on the fly and translates them into higher level SRL events, contingencies and graphs of users’ learning actions.

Findings of this research suggest that elements from both social and organizational aspects of a workplace should be integrated into the design and development of interventions which aim to support users’ SRL processes in that environment. Users’ perceived usefulness of the interventions show that they do consider the social context of their organization when planning their learning goals; yet, they prefer to know clearly what competences their organization expects them to achieve. Analysis of users’ trace data, on the other hand, indicates a relative balance between users’ reliance on both social and organizational contexts. The Social Wave intervention, which brought users updates from their social context, was the most central one during their learning actions, also the strongest determinant of users’ engagement in SRL processes. The next most central intervention included the one that informed users about how various learning resources were used by their colleagues, along with the two interventions providing users with the organizational context of their workplace.

This theoretically-grounded understanding can guide researchers in intervention planning and development for workplace settings. Also, the trace-based methodology developed within this work takes promising steps toward adopting new methodological approaches in investigating SRL, and offers new ways to achieve insight into factors that promote knowledge workers’ use of self-regulatory processes. Future research can gain substantially by applying social analytics on users’ trace data collected using trace methodologies, merged with other quantitative and qualitative means for gathering data about users’ SRL beliefs and processes.

Full Thesis | LinkedIn

• Adaptivity & Personalization

Karen Tanenbaum

User perceptions of adaptivity in ubiquitous systems: a critical exploration

Description: This dissertation addresses a gap in the field of designing adaptivity for ubiquitous systems by taking a critical look at the notion of "adaptivity" from the perspective of user experience. Through a set of detailed case studies of several different systems, I develop a set of concepts related to the experience of adaptivity. These concepts are supplemented by a set of design considerations that can assist in designers in thinking about key issues connected to the concepts. My work is a first take on untangling the complex relationship between ubiquity, adaptivity and the design of novel systems.

Through a collective case study, I examine the differences between the intended and actual experience of three adaptive systems: the Reading Glove, Kurio, and socio-echo. The Reading Glove was an interactive storytelling system involving a piece of wearable technology that allowed participants to trigger story information by picking up objects. An adaptive component guided the reader through the story by recommending objects to interact with next. Kurio was a museum guide system that involved playing an educational game distributed across a set of handheld and tabletop devices. The adaptive component attempted to gauge the appropriate learning level in assigning tasks to each individual. Socio-echo was a group game played in an ambient environment, where teams of players had to coordinate their physical movements to solve riddle-based levels. Characteristics of the group's movement, location and position were used to adapt the system's ambient feedback system.

From the analysis of these cases, I draw out a set of interrelated concepts that are useful for designing adaptive systems. The experience of adaptivity is impacted by the user's awareness of adaptivity and the interpretation of the adaptive effects. Factors like trust, surprise, augmentation, legibility, collapse, confusion, control and choice also play a role in grappling with intelligent components within complex systems. This research highlights the complexity involved in designing the adaptive components of computing systems making use of tangible and other novel interface styles by examining some of the experiential effects of these new interaction paradigms and how they relate to the intentions of the designers.

Full Thesis | LinkedIn

• Software Engineering

Samaneh Soltani

Towards automated feature model configuration with optimizing the non-functional requirements

Description: A Software Product Line is a family of software systems in a domain, which share some common features but also have significant variabilities. A feature model is a variability modeling artifact, which represents differences among software products with respect to the variability relationships among their features. Having a feature model along with a reference model developed in the domain engineering lifecycle, a concrete product of the family is derived by binding the variation points in the feature model (called configuration process) and by instantiating the reference model. However, feature model configuration is a cumbersome task because of: 1) the large number of features in industrial feature models, which increases the complexity of the configuration process; 2) the positive or negative impact of the features on non-functional properties; and 3) the stakeholders’ preferences with respect to the desirable non-functional properties of the final product. Several configuration techniques have already been proposed to facilitate automated product derivation. However, most of the current proposals are not designed to consider stakeholders’ preferences and constraints especially with regard to non-functional properties. In this work we address the feature model configuration problem and propose a framework, which employs an artificial intelligence planning technique to automatically select suitable features that satisfy both the functional and non-functional preferences and constraints of stakeholders. We also provide tooling support to facilitate the use of our framework. Our experiments show that despite the complexity involved in the simultaneous consideration of both functional and non-functional properties, our configuration technique is scalable.

Full Thesis | LinkedIn