Högskolan i Skövde

This chapter gives an overview of the content of this book, and links them with the work of Prof. Sadaaki Miyamoto, to whom this book is dedicated.

In this paper we discuss how three types of fuzzy partitions can be used to describe the results of three types of cluster structures. Standard fuzzy partitions are suitable for centroid based clusters, and I-fuzzy partitions for clusters represented by segments or lines (e.g., c-varieties). In this paper, we introduce hesitant fuzzy partitions. They are suitable for clusters defined by sets of centroids. Because of that, we show that they are useful for hierarchical clustering. We also establish the relationship between hesitant fuzzy partitions and I-fuzzy partitions.

The challenge of delivering personalized learning experiences is amplified by the size of classrooms and of online learning communities. In turn, serious games are increasingly recognized for their potential to improve education, but a typical requirement from instructors is to gain insight into how the students are playing. When we bring games into the rapidly growing online learning communities, the challenges multiply and hinder the potential effectiveness of serious games. There is a need to deliver a comprehensive, flexible and intelligent learning framework that facilitates better understanding of learners’ knowledge, effective assessment of their progress and continuous evaluation and optimization of the environments in which they learn. This paper aims to explore the potential in the use of games and learning analytics towards scaffolding and supporting teaching and learning experience. The conceptual model discussed aims to highlight key considerations that may advance the current state of learning analytics, adaptive learning and serious games, by leveraging serious games as an ideal medium for gathering data and performing adaptations. This opportunity has the potential to affect the design and deployment of education and training in the future.

We present a significantly improved implementation of a parallel SVM algorithm (PSVM) together with a comprehensive experimental study. Support Vector Machines (SVM) is one of the most well-known machine learning classification techniques. PSVM employs the Interior Point Method, which is a solver used for SVM problems that has a high potential of parallelism. We improve PSVM regarding its structure and memory management for contemporary processor architectures. We perform a number of experiments and study the impact of the reduced column size p and other important parameters as C and gamma on the class-prediction accuracy and training time. The experimental results show that there exists a threshold between the number of computational cores and the training time, and that choosing an appropriate value of p effects the choice of the C and gamma parameters as well as the accuracy.

The challenge of delivering personalized learning experiences is often increased by the size of classrooms and online learning communities. Serious Games (SGs) are increasingly recognized for their potential to improve education. However, the issues related to their development and their level of effectiveness can be seriously affected when brought too rapidly into growing online learning communities. Deeper insights into how the students are playing is needed to deliver a comprehensive and intelligent learning framework that facilitates better understanding of learners' knowledge, effective assessment of their progress and continuous evaluation and optimization of the environments in which they learn. This paper discusses current SOTA and aims to explore the potential in the use of games and learning analytics towards scaffolding and supporting teaching and learning experience. The conceptual model (ecosystem and architecture) discussed in this paper aims to highlight the key considerations that may advance the current state of learning analytics, adaptive learning and SGs, by leveraging SGs as an suitable medium for gathering data and performing adaptations.

Recently, innovative technology like Trackman has made it possible to generate data describing golf swings. In this application paper, we analyze Trackman data from 275 golfers using descriptive statistics and machine learning techniques. The overall goal is to find non-trivial and general patterns in the data that can be used to identify and explain what separates skilled golfers from poor. Experimental results show that random forest models, generated from Trackman data, were able to predict the handicap of a golfer, with a performance comparable to human experts. Based on interpretable predictive models, descriptive statistics and correlation analysis, the most distinguishing property of better golfers is their consistency. In addition, the analysis shows that better players have superior control of the club head at impact and generally hit the ball straighter. A very interesting finding is that better players also tend to swing flatter. Finally, an outright comparison between data describing the club head movement and ball flight data, indicates that a majority of golfers do not hit the ball solid enough for the basic golf theory to apply.

Golf is a popular sport around the world. Since an accomplished golf swing is essential for succeeding in this sport, golf players spend a considerable amount of time perfecting their swing. In order to guide the design of future computer-based training systems that support swing instruction, this paper analyzes the data gathered during interviews with golf instructors and participant observations of actual swing coaching sessions. Based on our field work, we describe the characteristics of a proficient swing, how the instructional sessions are normally carried out and the challenges professional instructors face. Taking into account these challenges, we outline which desirable capabilities future computer-based training systems for professional golf instructors should have.

We propose the use of evidential combination operators for advanced driver assistance systems (ADAS) for vehicles. More specifically, we elaborate on how three different operators, one precise and two imprecise, can be used for the purpose of entrapment prediction, i.e., to estimate when the relative positions and speeds of the surrounding vehicles can potentially become dangerous. We motivate the use of the imprecise operators by their ability to model uncertainty in the underlying sensor information and we provide an example that demonstrates the differences between the operators.

High performance is a goal in most sporting activities, for elite athletes as well as for recreational practitioners, and the process of measuring, evaluating and improving performance is one fundamental aspect to why people engage in sports. This is a complex process which possibly involves analyzing large amounts of heterogeneous data in order to apply actions that change important properties for improved outcome. The number of computer based decision support systems in the context of data analysis for performance improvement is scarce. In this position paper we briefly review the literature, and we propose the use of information fusion, situation modeling and visual analytics as suitable tools for supporting decision makers, ranging from recreational to elite, in the process of performance evaluation.