Biomedical subjects

Lawrence O Hall

Publications and source records attributed to Lawrence O Hall.

4 recordsLinked to original sources

PubMed2007

A comparison of decision tree ensemble creation techniques.

We experimentally evaluate bagging and seven other randomization-based approaches to creating an ensemble of decision tree classifiers. Statistical tests were performed on experimental results from 57 publicly available data sets. When cross-validation comparisons were tested for statistical significance, the best method was statistically more accurate than bagging on only eight of the 57 data sets. Alternatively, examining the average ranks of the algorithms across the group of data sets, we find that boosting, random forests, and randomized trees are statistically significantly better than bagging. Because our results suggest that using an appropriate ensemble size is important, we introduce an algorithm that decides when a sufficient number of classifiers has been created for an ensemble. Our algorithm uses the out-of-bag error estimate, and is shown to result in an accurate ensemble for those methods that incorporate bagging into the construction of the ensemble.

Algorithms↗

PubMed2004

Selection of patients for clinical trials: an interactive web-based system.

The purpose of a clinical trial is to evaluate a new treatment procedure. When medical researchers conduct a trial, they recruit participants with appropriate health problems and medical histories. To select participants, they analyze medical records of the available patients, which has traditionally been a manual procedure. We describe an expert system that helps to select patients for clinical trials. If the available data are insufficient for choosing patients, the system suggests additional medical tests and finds an ordering of the tests that reduces their total cost. Experiments show that the system can increase the number of selected patients. We also present an interface that enables a medical researcher to add clinical trials and selection criteria without the help of a programmer. The addition of a new trial takes 10-20 min, and novice users learn the functionality of the interface in about an hour.

Breast Neoplasms↗

PubMed2004

Recognizing plankton images from the shadow image particle profiling evaluation recorder.

We present a system to recognize underwater plankton images from the shadow image particle profiling evaluation recorder (SIPPER). The challenge of the SIPPER image set is that many images do not have clear contours. To address that, shape features that do not heavily depend on contour information were developed. A soft margin support vector machine (SVM) was used as the classifier. We developed a way to assign probability after multiclass SVM classification. Our approach achieved approximately 90% accuracy on a collection of plankton images. On another larger image set containing manually unidentifiable particles, it also provided 75.6% overall accuracy. The proposed approach was statistically significantly more accurate on the two data sets than a C4.5 decision tree and a cascade correlation neural network. The single SVM significantly outperformed ensembles of decision trees created by bagging and random forests on the smaller data set and was slightly better on the other data set. The 15-feature subset produced by our feature selection approach provided slightly better accuracy than using all 29 features. Our probability model gave us a reasonable rejection curve on the larger data set.

Algorithms↗

PubMed2004

Comments on "a parallel mixture of SVMs for very large scale problems".

Collobert, Bengio, and Bengio (2002) recently introduced a novel approach to using a neural network to provide a class prediction from an ensemble of support vector machines (SVMs). This approach has the advantage that the required computation scales well to very large data sets. Experiments on the Forest Cover data set show that this parallel mixture is more accurate than a single SVM, with 90.72% accuracy reported on an independent test set. Although this accuracy is impressive, their article does not consider alternative types of classifiers. We show that a simple ensemble of decision trees results in a higher accuracy, 94.75%, and is computationally efficient. This result is somewhat surprising and illustrates the general value of experimental comparisons using different types of classifiers.

Algorithms↗