Computer Science Technical Reports: Recent submissions
Now showing items 21-40 of 280
-
Feature Extraction for Dynamic Integration of Classifiers
(Trinity College Dublin, Department of Computer Science, 2006)Recent research has shown the integration of multiple classifiers to be one of the most important directions in machine learning and data mining. In this paper, we present an algorithm for the dynamic integration of ... -
The Best Way to Instil Confidence is by Being Right An Evaluation of the Effectiveness of Case-Based Explanations in providing User Confidence
(Trinity College Dublin, Department of Computer Science, 2005-02-07)Instilling confidence in the abilities of machine learning systems in end-users is seen as critical to their success in real world problems. One way in which this can be achieved is by providing users with interpretable ... -
Dynamic Integration of Classifiers for Tracking Concept Drift in Antibiotic Resistance Data
(Trinity College Dublin, Department of Computer Science, 2005-02-17)In the real world concepts are often not stable but change with time. A typical example of this in the medical context is antibiotic resistance, where pathogen sensitivity may change over time as new pathogen strains ... -
Re-using Implicit Knowledge in Short-term Information Profiles for Context-sensitive Tasks
(Trinity College Dublin, Department of Computer Science, 2005-02-22)Typically, case-based recommender systems recommend single items to the on-line customer. In this paper we introduce the idea of recommending a user-defined collection of items where the user has implicitly encoded ... -
Blame-Based Noise Reduction: An Alternative Perspective on Noise Reduction for Lazy Learning
(Trinity College Dublin, Department of Computer Science, 2005-02-22)In this paper we present a new perspective on noise reduction for nearest-neighbour classifiers. Classic noise reduction algorithms such as Repeated Edited Nearest Neighbour remove cases from the training set if they are ... -
Meta-Knowledge Management in MultiStrategy Process-Oriented Knowledge Discovery Systems
(Trinity College Dublin, Department of Computer Science, 2005) -
On the use of Information Systems Research Methods in Data Mining
(Trinity College Dublin, Department of Computer Science, 2005) -
Using Early-Stopping to Avoid Overfitting in Wrapper-Based Feature Selection Employing Stochastic Search
(Trinity College Dublin, Department of Computer Science, 2005-05-11)It is acknowledged that overfitting can occur in feature selection using the wrapper method when there is a limited amount of training data available. It has also been shown that the severity of overfitting is related ... -
Sequential Genetic Search for Ensemble Feature Selection
(Trinity College Dublin, Department of Computer Science, 2005)Ensemble learning constitutes one of the main directions in machine learning and data mining. Ensembles allow us to achieve higher accuracy, which is often not achievable with single models. One technique, which proved to ... -
Knowledge Discovery in Microbiology Data: Analysis of Antibiotic Resistance in Nosocomial Infections
(Trinity College Dublin, Department of Computer Science, 2005)The goal of this paper is to address the currently serious problem of antibiotic resistance applying knowledge discovery techniques to real hospital data. In this paper we introduce our approach to that problem and the ... -
Using Early Stopping to Reduce Overfitting in Wrapper-Based Feature Weighting
(Trinity College Dublin, Department of Computer Science, 2005-05-18)It is acknowledged that overfitting can occur in feature selection using the wrapper method when there is a limited amount of training data available. It has also been shown that the severity of overfitting is related to ... -
Producing Accurate Interpretable Clusters from High-Dimensional Data
(Trinity College Dublin, Department of Computer Science, 2005-05-19)The primary goal of cluster analysis is to produce clusters that accurately reflect the natural groupings in the data. A second objective that is important for high-dimensional data is to identify features that are ... -
A Review of Active Learning and Co-Training in Text Classification
(Trinity College Dublin, Department of Computer Science, 2005-11-04) -
The Benefits of Using a Complete Probability Distribution when Decision Making: An Example in Anticoagulant Drug Therapy
(Trinity College Dublin, Department of Computer Science, 2005-08-06)In this paper we aim to show how probabilistic prediction of a continuous variable could be more beneficial to a medical practitioner than classification or numeric/point prediction of the same variable in many scenarios. ... -
SAMPLE: An On-Demand Probabilistic Routing Protocol for Ad-hoc Networks
(Trinity College Dublin, Department of Computer Science, 2004-01-21)Existing on-demand ad hoc routing protocols assume an idealised wireless network in which all links in the network are either on or off and where all functioning links are equally good. Such a model interprets the fraction ... -
Where's Waldo? - or - A taxonomy for thinking about location in pervasive computing
(Trinity College Dublin, Department of Computer Science, 2004-02-06)[Introduction] Virtually all pervasive computing systems use some form of location for affecting the system's behaviour. Location-based services are available commercially, albeit in a primitive form, from many mobile ... -
Random subspacing for regression ensembles
(Trinity College Dublin, Department of Computer Science, 2004)In this work we present a novel approach to ensemble learning for regression models, by combining the ensemble generation technique of random subspace method with the ensemble integration methods of Stacked Regression ... -
Dynamic Integration of Regression Models
(Trinity College Dublin, Department of Computer Science, 2004-03-26)In this paper we adapt the recently proposed Dynamic Integration ensemble techniques for regression problems and compare their performance to the base models and to the popular ensemble technique of Stacked Regression. We ... -
An Analysis of Case-Base Editing in a Spam Filtering System
(Trinity College Dublin, Department of Computer Science, 2004-08)Because of the volume of spam email and its evolving nature, any deployed Machine Learning-based spam filtering system will need to have procedures for case-base maintenance. Key to this will be procedures to edit ... -
A Case-Based Explanation System for `Black-Box? Systems
(Trinity College Dublin, Department of Computer Science, 2004-06-23)Most users of machine-learning products are reluctant to use the systems without any sense of the underlying logic that has led to the system?s predictions. Unfortunately many of these systems lack any transparency in the ...
