Splitting methods for decision tree induction: An exploration of the relative performance of two entropy-based families

Osei-Bryson, Kweku-Muata; Giles, Kendall

doi:10.1007/s10796-006-8779-8

Cited by 9 publications

(2 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For KNN and decision tree, the best value of parameter k and cp were decided through attempting different values. In addition, the performance of two split methods, information gain and Gini impurity were also investigated for decision tree [101]. The value corresponding to the highest accuracy (10-fold cross validation with 3 times repetition) was selected as the optimal values for the parameters.…”

Section: Classificationmentioning

confidence: 99%

A Pervasive Sensing Approach to Automatic Assessment of Trunk Coordination Using Mobile Devices

Liang¹,

Yoshida²,

Iino³

et al. 2018

EAI Endorsed Trans Perv Health Tech

View full text Add to dashboard Cite

Assessing trunk coordination has many potential applications in health promotion. However, traditional bio-mechanical approaches are not suited for daily use as they require expensive devices and manual analysis. This study aimed to develop an approach for automatic classification of good and poor trunk coordination using widely available mobile devices. We investigated different combinations of sensor locations (i.e. chest and pelvis), sensing modalities (i.e. accelerometer and gyroscope) and classification techniques (i.e. SVM, KNN, and decision tree). Results showed that using both sensing modalities at chest and pelvis with SVM produced the best classification accuracy: 96% for chest rotation and 100% for pelvis rotation. In practice, however, using one device with both sensing modalities (i.e. accelerometer and gyroscope) will achieve a better trade-off between feasibility and accuracy. In this case, the device should be fixed on the chest. KNN should be selected as the classification technique for chest rotation (best accuracy 95%), and SVM should be selected as the classification technique for pelvis rotation (best accuracy 79%). Post hoc analysis found that poor coordination during chest rotation was associated to weak cross-correlation of angular velocity between chest and pelvis in the frontal plane, while poor coordination during pelvis rotation was associated to weak correlations of angular velocity between the three orthogonal components at chest. This study demonstrated how simple mobile devices can capture relevant motion data and extract key features that help construct computational models for automatic assessment of trunk coordination.

show abstract

Section: Classificationmentioning

confidence: 99%

A Pervasive Sensing Approach to Automatic Assessment of Trunk Coordination Using Mobile Devices

Liang¹,

Yoshida²,

Iino³

et al. 2018

EAI Endorsed Trans Perv Health Tech

View full text Add to dashboard Cite

show abstract

“…However, most of these investigations have focussed on classification rather than regression trees, where the popular alternatives are chi-squared, entropy and posterior improvement criteria (Taylor and Silverman 1993;Shi 1999;Osei-Bryson 2006). One study that does include regression trees is that by Bremner and Ross (2002), which reports greater success in modelling data that include interaction effects by local averaging of residual sum of squares, but for our purposes none of these investigations suggests a change of focus from the W criterion.…”

Section: Introductionmentioning

confidence: 99%

A recursive partitioning tool for interval prediction

Krzanowski

Hand

2007

ADAC

View full text Add to dashboard Cite

The traditional approach to regression trees involves partitioning the space of predictor variables into subsets that optimise a function of the response variable(s), and then predicting future response values by a single-valued summary statistic in each subset. Our belief is that a prediction interval is of greater practical use than a predictive value, and that the criterion for the partitioning should be based on such intervals rather than on single values. We define four potential criteria in the case of a single response variable, discuss computational aspects of producing the partition, evaluate the criteria on both real and simulated data, and draw some tentative conclusions about their relative efficacies. The methodology is extended to the case of multiple response variables, and its viability is demonstrated by application to some further real data. The possibility of fitting distributions to within-subsets data is discussed, and some potential extensions are briefly outlined.

show abstract

Comparing data mining methods with logistic regression in childhood obesity prediction

et al. 2009

View full text Add to dashboard Cite

The epidemiological question of concern here is "can young children at risk of obesity be identified from their early growth records?" Pilot work using logistic regression to predict overweight and obese children demonstrated relatively limited success. Hence we investigate the incorporation of non-linear interactions to help improve accuracy of prediction; by comparing the result of logistic regression with those of six mature data mining techniques.The contributions of this paper are as follows: a) a comparison of logistic regression with six data mining techniques: specifically, for the prediction of overweight and obese children at 3 years using data recorded at birth, 6 weeks, 8 months and 2 years respectively; b) improved accuracy of prediction: prediction at 8 months accuracy is improved very slightly, in this case by using neural networks, whereas for prediction at 2 years obtained accuracy is improved by over 10%, in this case by using Bayesian methods. It has also been shown that incorporation of non-linear interactions could be important in epidemiological prediction, and that data mining techniques are becoming sufficiently well established to offer the medical research community a valid alternative to logistic regression.

show abstract

Splitting methods for decision tree induction: An exploration of the relative performance of two entropy-based families

Cited by 9 publications

References 12 publications

A Pervasive Sensing Approach to Automatic Assessment of Trunk Coordination Using Mobile Devices

A Pervasive Sensing Approach to Automatic Assessment of Trunk Coordination Using Mobile Devices

A recursive partitioning tool for interval prediction

Comparing data mining methods with logistic regression in childhood obesity prediction

Contact Info

Product

Resources

About