nach oben

2016 | Buch

Learning Analytics in R with SNA, LSA, and MPIA

verfasst von: Fridolin Wild

Verlag: Springer International Publishing

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

This book introduces Meaningful Purposive Interaction Analysis (MPIA) theory, which combines social network analysis (SNA) with latent semantic analysis (LSA) to help create and analyse a meaningful learning landscape from the digital traces left by a learning community in the co-construction of knowledge.

The hybrid algorithm is implemented in the statistical programming language and environment R, introducing packages which capture – through matrix algebra – elements of learners’ work with more knowledgeable others and resourceful content artefacts. The book provides comprehensive package-by-package application examples, and code samples that guide the reader through the MPIA model to show how the MPIA landscape can be constructed and the learner’s journey mapped and analysed. This building block application will allow the reader to progress to using and building analytics to guide students and support decision-making in learning.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Introduction

Abstract

This book presents new ways to automatically represent conceptual development evident from interaction of learners with content. It introduces novel instruments for further analysis and visualization, therewith supporting guidance and decision-making about and during learning. Within this introduction, the scope of the book is further refined. An introduction to the learning technology software market is followed by a characterization of the research challenge faced and of the wider research area in which the solution presented in this book is situated. Three roots of the work are introduced in this chapter to set the stage: epistemic theory, algorithms for representation/visualization/analysis, and, finally, the application area of Learning Analytics. The chapter is rounded up by an overview over the book.

Fridolin Wild

Chapter 2. Learning Theory and Algorithmic Quality Characteristics

Abstract

This chapter provides the theoretical foundations from which conceptual models for real learning analytics systems can be derived. It will lead over to the one algorithmic model created with the help of this theory and further detailed in this book, a novel algorithm called ‘meaningful purposive interaction analysis’. Conceptually, it is grounded in methodical culturalism and its novelty is to extend this culturalist information theory with the constructs ‘learning’, ‘competence’, and ‘performance’. This way, novel theory is established: culturalist learning theory—as an extension of culturalist information theory.

Fridolin Wild

Chapter 3. Representing and Analysing Purposiveness with SNA

Abstract

Social Network Analysis is a powerful instrument for representing and analysing networks and the relational structure expressed in the incidences from which they are constructed. This chapter provides an introduction (including data preparation) to the available analytical instruments. Both a foundational and an extended application example demonstrate how (social) network analysis is applied in practice.

Fridolin Wild

Chapter 4. Representing and Analysing Meaning with LSA

Abstract

Latent Semantic Analysis (LSA) is a time-tested algorithm for representing and analysing meaning from text, with its closeness in mathematical foundation being a natural candidate for further integration. The chapter starts with its mathematical foundations, then provides an overview on the standard analysis workflow (with the package developed), also guiding through the standard use cases. Two demos follow subsequently to foster understanding and gain insight into the main restrictions applying to LSA. The foundational example presented picks up the usage scenario of the foundational SNA demo of the previous chapter. Following a summary of the state of the art in application of LSA to technology-enhanced learning, a second, real-life application example in essay scoring will be added. A summary outlining also the key limitations of LSA concludes the chapter.

Fridolin Wild

Chapter 5. Meaningful, Purposive Interaction Analysis

Abstract

This chapter introduces to the mathematical foundation of MPIA in matrix theory. The shared mathematical foundation with (social) network analysis and latent semantic analysis will be explored in order to make visible where MPIA goes beyond the sum of these predating methods. It therefore, first, revisits essential theorems about matrix algebra, as a side effect introducing also the basic mathematic vocabulary used in this section. Then, Eigensystem calculation and singular value decomposition are deduced, adding detail that will serve understanding of the MPIA analysis, visualisation, and inspection methods that make use of it. The proximity and identity relations are at the core of this and consequently their working principles and foundations are discussed subsequently. Working examples illustrate how the matrix theory foundations (and the higher level transformations making use of it) relate to the incidence matrices of the social network analysis and latent semantic analysis constituents.

Fridolin Wild

Chapter 6. Visual Analytics Using Vector Maps as Projection Surfaces

Abstract

This chapter introduces geometrical projection surfaces to provide a stable ‘stage’ for subsequent visual analysis of locations, positions, and pathways. It proposes methods for link erosion, planar projection (with monotonic convergence!), kernel smoothening, and a hypsometric colour scheme for tile colouring in order to help create conceptual landscape visualizations. The means presented here complement the analytical processes introduced in the previous chapter with a powerful visual instrument. When combined, the two provide means to analyse social semantic performance networks.

Fridolin Wild

Chapter 7. Calibrating for Specific Domains

Abstract

Eigenspace-based models were shown to exhibit greater effectiveness than their simple vector space counterparts in settings that benefit from fuzziness (such as information retrieval or recommender systems). In settings that require precision in representation structure (such as in essay scoring or for conceptual relationship mining), however, improved means to predict model behaviour from parameter settings could ease applicability and increase efficiency by reducing tuning times.

This chapter reports experiences and experiment results from a systematic investigation of tuning parameters, their potential settings, and interdependencies between them. This includes studying the influence of sanitising operations, sampling, dimensionality changes, and degrees of specialisation. Trends indicate that the smaller the corpus, the more domain-specific documents are required. Moreover, recommendations for vocabulary filtering can be derived, dependent on the size of the corpus.

Fridolin Wild

Chapter 8. Implementation: The MPIA Package

Abstract

This chapter provides an introduction to the actual implementation, i.e. the mpia R package, its class system as well as key data manipulation and visualization methods. It provides the instruments to automatically represent and further analyse the conceptual development evident from the performance of learners, while at the same time introducing re-representation mechanisms for visual analytics to guide and support decision-making about and during learning. Use case and activity diagrams of an idealized analysis workflow foster understanding of how MPIA analysis processes can be conducted.

Fridolin Wild

Chapter 9. MPIA in Action: Example Learning Analytics

Abstract

This chapter presents comprehensive application examples of MPIA, thereby contributing an extended, more holistic approach to learning analytics, as the included review of the state of the art will show. Already the previous chapter illustrated the general analysis workflow with the mpia software package with the help of a parsimonious-code example. This chapter now aims to extend this and showcase MPIA inaction, using increasingly more complex learning analytics applications and—in the two final ones—real-life data.

Fridolin Wild

Chapter 10. Evaluation

Abstract

The research conducted for this book is accompanied by a series of evaluation studies, each of them contributing to both conceptual as well as algorithmic refinement and—finally—to their concluding assessment. In this chapter, a precise summary of these studies is synthesised. The studies in sum, extended by a final accuracy study, provide evidence of the validity and utility of the approach developed.

Fridolin Wild

Chapter 11. Conclusion and Outlook

Abstract

This final chapter summarises the achievements against the problem statement introduced in the first chapter, critically examining the contributions along the derived three objectives. The three derived research objectives are revisited, relating them with the contributions brought forward in this book. Furthermore, connections to other research areas are pointed out and open points for future research defined.

Fridolin Wild

Erratum to: Learning Analytics in R with SNA, LSA, and MPIA

Fridolin Wild

Backmatter

Titel: Learning Analytics in R with SNA, LSA, and MPIA
verfasst von: Fridolin Wild
Verlag: Springer International Publishing
Electronic ISBN: 978-3-319-28791-1
Print ISBN: 978-3-319-28789-8
DOI: https://doi.org/10.1007/978-3-319-28791-1