Classification Trees for Ordinal Responses in<i>R</i>: The<b>rpartScore</b>Package

Galimberti, Giuliano; Soffritti, Gabriele; Maso, Matteo Di

doi:10.18637/jss.v047.i10

Cited by 44 publications

(42 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For illustrative purposes, we applied a support vector machine (SVM) algorithm and two ordinal classification algorithms that were available in R to the dataset. The SVM model, which uses hyperplanes to optimize the linear separation between classes, was implemented using the e1071 package [25] in R. The two ordinal prediction models were constructed using the R packages glmnetcr [14] and rpartScore [16] . In glmnetcr, ordinal response data is modeled with an L 1 penalized continuation ratio model.…”

Section: Resultsmentioning

confidence: 99%

Cost-sensitive performance metric for comparing multiple ordinal classifiers

George

Chang

2016

AIR

View full text Add to dashboard Cite

The surge of interest in personalized and precision medicine during recent years has increased the application of ordinal classification problems in biomedical science. Currently, accuracy, Kendall’s τb, and average mean absolute error are three commonly used metrics for evaluating the effectiveness of an ordinal classifier. Although there are benefits to each, no single metric considers the benefits of predictive accuracy with the tradeoffs of misclassification cost. In addition, decision analysis that considers pairwise analysis of the metrics is not trivial due to inconsistent findings. A new cost-sensitive metric is proposed to find the optimal tradeoff between the two most critical performance measures of a classification task – accuracy and cost. The proposed method accounts for an inherent ordinal data structure, total misclassification cost of a classifier, and imbalanced class distribution. The strengths of the new methodology are demonstrated through analyses of three real cancer datasets and four simulation studies. The new cost-sensitive metric proved better performance in its ability to identify the best ordinal classifier for a given analysis. The performance metric devised in this study provides a comprehensive tool for comparative analysis of multiple (and competing) ordinal classifiers. Consideration of the tradeoff between accuracy and misclassification cost in decisions regarding ordinal classification problems is imperative in real-world application. The work presented here is a precursor to the possibility of incorporating the proposed metric into a prediction modeling algorithm for ordinal data as a means of integrating misclassification cost in final model selection.

show abstract

Section: Resultsmentioning

confidence: 99%

Cost-sensitive performance metric for comparing multiple ordinal classifiers

George

Chang

2016

AIR

View full text Add to dashboard Cite

show abstract

“…We employed the “rpartScore” package in R to build the decision tree (DT) models[8]. The package provides functions to build classification trees for ordinal responses within the classification and regression tree (CART) framework.…”

Section: Methodsmentioning

confidence: 99%

Predictive modeling for classification of positive valence system symptom severity from initial psychiatric evaluation records

Posada

Barda

Shi

et al. 2017

Journal of Biomedical Informatics

View full text Add to dashboard Cite

In response to the challenges set forth by the CEGS N-GRID 2016 Shared Task in Clinical Natural Language Processing, we describe a framework to automatically classify initial psychiatric evaluation records to one of four positive valence system severities: absent, mild, moderate, or severe. We used a dataset provided by the event organizers to develop a framework comprised of natural language processing (NLP) modules and 3 predictive models (two decision tree models and one Bayesian network model) used in the competition. We also developed two additional predictive models for comparison purpose. To evaluate our framework, we employed a blind test dataset provided by the 2016 CEGS N-GRID. The predictive scores, measured by the macro averaged-inverse normalized mean absolute error score, from the two decision trees and Naïve Bayes models were 82.56%, 82.18%, and 80.56%, respectively. The proposed framework in this paper can potentially be applied to other predictive tasks for processing initial psychiatric evaluation records, such as predicting 30-day psychiatric readmissions.

show abstract

“…An ordinal response, however, suggests that the cost of misClassification should increase with | j − k |. Although there are alternative ways to incorporate category specific misClassification costs [8, 12, 25], we adopt the generalized Gini index with either the absolute cost function, C ( j | k ) = | j − k |, or the quadratic cost function, C ( j | k )= ( j − k ) 2 , as implemented in rpartScore [12]. …”

Section: Definitions and Notationmentioning

confidence: 99%

“…The rpartScore package [12] is used to fit an ordinal CART model using the generalized Gini index with the absolute cost function. CART models fit with rpartScore are pruned using the default pruning parameters provided in the package.…”

Section: Comparison Of Olr and Cartmentioning

confidence: 99%

“…A common approach for identifying optimal splits uses the Gini impurity index (available in the rpart package in R), a measure that associates the same cost with all misclassifications [12]. Alternatively, a generalized Gini index (available in the rpartScore package in R [12]) allows the misclassification cost to increase as the distance between the true and predicted response category increases [5, 8, 12, 25]. Therefore, while the Gini impurity index is appropriate as a splitting criterion for multinomial response data, the generalized Gini index is more suitable for ordinal data.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Ordinal Logic Regression: A classifier for discovering combinations of binary markers for ordinal outcomes

Wolf

Slate

Hill

2015

Computational Statistics & Data Analysis

View full text Add to dashboard Cite

In medicine, it is often useful to stratify patients according to disease risk, severity, or response to therapy. Since many diseases arise from complex gene-gene and gene-environment interactions, patient strata may be defined by combinations of genetic and environmental factors. Traditional statistical methods require specifying interactions a priori making it difficult to identify high order interactions. Alternatively, machine learning methods can model complex interactions, however these models are often difficult to interpret in a clinical setting. Logic regression (LR) enables modeling a binary outcome using logical combinations of binary predictors yielding easily interpretable models. However LR, as currently available, cannot model ordinal responses. This paper extends LR to model an ordinal response and the resulting method is called Ordinal Logic Regression (OLR). Several simulations comparing OLR and Classification and Regression Trees (CART) demonstrate that OLR is superior to CART for identifying variable interactions associated with an ordinal response. OLR is applied to data from a study to determine associations between genetic and health factors with severity of adult periodontitis.

show abstract

Classification Trees for Ordinal Responses inR: TherpartScorePackage

Cited by 44 publications

References 20 publications

Cost-sensitive performance metric for comparing multiple ordinal classifiers

Cost-sensitive performance metric for comparing multiple ordinal classifiers

Predictive modeling for classification of positive valence system symptom severity from initial psychiatric evaluation records

Ordinal Logic Regression: A classifier for discovering combinations of binary markers for ordinal outcomes

Contact Info

Product

Resources

About