Showing papers by "Mark Hall published in 2001"

A Simple Approach to Ordinal Classification

Abstract: Machine learning methods for classification problems commonly assume that the class values are unordered. However, in many practical applications the class values do exhibit a natural order--for example, when learning how to grade. The standard approach to ordinal classification converts the class value into a numeric quantity and applies a regression learner to the transformed data, translating the output back into a discrete class value in a post-processing step. A disadvantage of this method is that it can only be applied in conjunction with a regression scheme. In this paper we present a simple method that enables standard classification algorithms to make use of ordering information in class attributes. By applying it in conjunction with a decision tree learner we show that it outperforms the naive approach, which treats the class values as an unordered set. Compared to special-purpose algorithms for ordinal classification our method has the advantage that it can be applied without any modification to the underlying learning scheme.

A cost of quality analysis of a building project: towards a complete methodology for design and build

Abstract: A number of studies have been published that claim to carry out cost of quality (COQ) studies on construction projects. These studies, however, have largely ignored the contribution of prevention and appraisal costs to COQ, and have limited their analysis to the impact of quality failures on the main contractor. This paper presents a methodology for assessing the ‘complete’ COQ for construction projects and reports on the findings of a building project in the UK on which the methodology was piloted. The company that applied this approach has now extended it to other projects.

Determining Progression in Glaucoma Using Visual Fields

Abstract: The standardized visual field assessment, which measures visual function in 76 locations of the central visual area, is an important diagnostic tool in the treatment of the eye disease glaucoma. It helps determine whether the disease is stable or progressing towards blindness, with important implications for treatment. Automatic techniques to classify patients based on this assessment have had limited success, primarily due to the high variability of individual visual field measurements. The purpose of this paper is to describe the problem of visual field classification to the data mining community, and assess the success of data mining techniques on it. Preliminary results show that machine learning methods rival existing techniques for predicting whether glaucoma is progressing--though we have not yet been able to demonstrate improvements that are statistically significant. It is likely that further improvement is possible, and we encourage others to work on this important practical data mining problem.

Root Cause Analysis - a Tool for Closer Supply Chain Integration in Construction

Abstract: Perhaps more than most industries, successful construction in the UK relies on effectively bringing together a diverse range of suppliers in what is often referred to as a ‘temporary-multi organization’ Relationships and integration between the organizations within this diverse and complex supply chain are difficult to form and maintain Frequently, relationships degenerate and become acrimonious, with companies attempting to protect their own position to the detriment of the construction project This paper outlines the root cause analysis tool and shows, using real-life case studies, how it can be used not only to identify and resolve problems but, at the same time engender better relationships and understanding between suppliers of services and products for a construction project

Experiences with a weighted decision tree learner

Abstract: Machine learning algorithms for inferring decision trees typically choose a single “best” tree to describe the training data. Recent research has shown that classification performance can be significantly improved by voting predictions of multiple, independently produced decision trees. This paper describes an algorithm, OB1, that produces a weighted sum over many possible models. Model weights are determined by the prior probability of the model, as well as the performance of the model during training. We describe an implementation of OBI that includes all possible decision trees as well as naive Bayesian models within a single option tree. Constructing all possible decision trees is very expensive, growing exponentially in the number of attributes. However it is possible to use the internal structure of the option tree to avoid recomputing values. In addition, the current implementation allows the option tree to be depth bounded.