New kernel functions and learning methods for text and data mining

Recent advances in machine learning methods enable increasingly the automatic construction of various types of computer assisted methods that have been difficult or laborious to program by human experts. The tasks for which this kind of tools are needed arise in many areas, here especially in the fields of bioinformatics and natural language processing. The machine learning methods may not work satisfactorily if they are not appropriately tailored to the task in question. However, their learning performance can often be improved by taking advantage of deeper insight of the application domain or the learning problem at hand. This thesis considers developing kernel-based learning algorithms incorporating this kind of prior knowledge of the task in question in an advantageous way. Moreover, computationally efficient algorithms for training the learning machines for specific tasks are presented. In the context of kernel-based learning methods, the incorporation of prior knowledge is often done by designing appropriate kernel functions. Another well-known way is to develop cost functions that fit to the task under consideration. For disambiguation tasks in natural language, we develop kernel functions that take account of the positional information and the mutual similarities of words. It is shown that the use of this information significantly improves the disambiguation performance of the learning machine. Further, we design a new cost function that is better suitable for the task of information retrieval and for more general ranking problems than the cost functions designed for regression and classification. We also consider other applications of the kernel-based learning algorithms such as text categorization, and pattern recognition in differential display. We develop computationally efficient algorithms for training the considered learning machines with the proposed kernel functions. We also design a fast cross-validation algorithm for regularized least-squares type of learning algorithm. Further, an efficient version of the regularized least-squares algorithm that can be used together with the new cost function for preference learning and ranking tasks is proposed. In summary, we demonstrate that the incorporation of prior knowledge is possible and beneficial, and novel advanced kernels and cost functions can be used in algorithms efficiently.

CRM and higher education : developing a monitoring system to improve relationships in e-learning environments

Peer-reviewed

Broadening the CRM domain of application to higher educational (e-learning) environments

In this paper, we reflect about the broadening of the field of application of CRM from the business domain to a wider context of relationships in which the inclusion of non-profit making organizations seems natural. In particular, we focus on analyzing the suitability of adopting CRM processes by universities and higher educational institutions dedicated to e-learning. This is an issue that, in our opinion, has much potential but has received little attention in research so far.

Enriching e-learning metadata through digital library usage analysis

Peer-reviewed

University teacher competences in a virtual teaching/learning environment : analysis of a teacher training experience

This paper attempts to shed light on the competencies a teacher must have inorder to teach in online university environments. We will relate a teacher trainingexperience, which was designed taking into account the methodological criteriaestablished in line with previous theoretical principles. The main objective of ouranalysis is to identify the achievements and difficulties of a specific formativeexperience, with the ultimate goal of assessing the suitability of this conceptualmethodologicalframework for the design of formative proposals aiming to contribute tothe development of teacher competencies for virtual environments.

Towards automated specifications of scenarios in enhanced learning technology

Recent standardization efforts in e-learning technology have resulted in a number of specifications, however, the automation process that is considered essential in a learning management system (LMS) is a lessexplored one. As learning technology becomes more widespread and more heterogeneous, there is a growing need to specify processes that cross the boundaries of a single LMS or learning resource repository. This article proposes to obtain a specification orientated to automation that takes on board the heterogeneity of systems and formats and provides a language for specifying complex and generic interactions. Having this goal in mind, a technique based on three steps is suggested. The semantic conformance profiles, the business process management (BPM) diagram, and its translation into the business process execution language (BPEL) seem to be suitable for achieving it.

User Centered Desing in E-Learning Environments : from Usability to Learner Experience

The advent of the Internet had a great impact on distance education and rapidly e-learning has become a killer application. Education institutions worldwide are taking advantage of the available technology in order to facilitate education to a growing audience. Everyday, more and more people use e-learning systems, environments and contents for both training and learning. E-learning promotes educationamong people that due to different reasons could not have access to education: people who could nottravel, people with very little free time, or withdisabilities, etc. As e-learning systems grow and more people are accessing them, it is necessary to consider when designing virtual environments the diverse needs and characteristics that different users have. This allows building systems that people can use easily, efficiently and effectively, where the learning process leads to a good user experience and becomes a good learning experience.

Capturing user behavior in e-learning environments

Peer-reviewed

Analyzing lifelong learning student behavior in a progressive degree

Peer-reviewed

The practical application of e-Portfolio at the Open University of Catalonia : assessment of competence based learning

The paper presents the results of the piloting or pilot test in a virtual classroom. This e-portfolio was carried out in the 2005-2006 academic year, with students of the Doctorate in Information Society, at the Open University of Catalonia. The electronic portfolio is a strategy for competence based assessment. This experience shows the types of e-portfolios, where students show their work without interactions, and apply the competence-based learning theories in an interactive portfolio system. The real process of learning is developed in the competency based system, the portfolio not only is a basic bio document, has become a real space for learning with competence model. The paper brings out new ideas and possibilities: the competence-based learning promotes closer relationships between universities and companies and redesigns the pedagogic act.

A Sparse Bayesian Approach for Joint Feature Selection and Classifier Learning

Peer-reviewed

A conceptual model for grid learning services automatic composition

Peer-reviewed

Collaborative group membership and access control for JXTA

Peer-reviewed

Towards the Construction of a Multi-agent Approach for Discovering the Meaning of Natural Language Collaborative Conversations

Peer-reviewed