Feature selection using Bayesian and multiclass Support Vector Machines approaches: Application to bank risk prediction

Feki, Asma; Ishak, Anis Ben; Feki, Saber

doi:10.1016/j.eswa.2011.08.172

Cited by 29 publications

(17 citation statements)

References 27 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Indeed, all the curves depicted in the top panels display the expected typical behavior; decreasing to reach a global minimum then increasing. This typical behavior reflects good performance of the stepwise algorithm and jointly attests good quality of the variable ranking [22,38]. This typical behavior is far from being realized with the SVR model especially when using the RFS hierarchy.…”

Section: Experiments On the Training Setsmentioning

confidence: 99%

Ozone Concentration Forecasting Using Statistical Learning Approaches

Ishak¹,

Daoud²,

Trabelsi³

2017

JMES

View full text Add to dashboard Cite

In this paper, we are interested in the statistical modeling and forecasting of the daily maximum ozone concentration in three monitoring stations from Tunisia. A large number of explicative variables has been considered in our study. We have focused our attention on the problem of variable selection in order to improve the forecasting performance. To achieve our goal, we have used Support Vector Regression (SVR) and Random Forests (RF). The main novelties of this paper are: the variety and originality of the approaches for variable selection in regression, and the audaciousness to deal with a sticky situation characterized by a relatively big pannier of explicative variables compared to the number of observations. The experimental results demonstrate that Random Forests outperform Support Vector Regression in variable ranking and selection. Finally, it was shown that the forecasting accuracy is at least preserved, for the three stations, when using only the selected variables.

show abstract

Section: Experiments On the Training Setsmentioning

confidence: 99%

Ozone Concentration Forecasting Using Statistical Learning Approaches

Ishak¹,

Daoud²,

Trabelsi³

2017

JMES

View full text Add to dashboard Cite

show abstract

“…This typical behavior reflects good performance of the stepwise algorithm and jointly attests good quality of the variables ranking. This behavior was deeply analyzed in the work of Ghattas and Ben Ishak (2008) for binary classification and more recently in the work of Feki et al (2012) for multiclass problems. Moreover, the OOB curve (OOB-MSE) seems a little bit optimistic than the RS curve (RS-MSE) obtained by random splitting for RF.…”

Section: Stepwise Curve Shapementioning

confidence: 99%

“…The set of variables leading to the model of smallest error rate is selected. Our stepwise algorithm has shown a promising results on classification problems even in the situations exposing the curse of dimensionality phenomenon, i.e., when the number of input variables p is very large compared to the sample size n (Ghattas and Ben Ishak, 2008;Feki et al, 2012).…”

Section: Variable Selection Algorithmsmentioning

confidence: 99%

Variable Selection Based on Statistical Learning Approaches to Improve PM10 Concentration Forecasting

Ishak¹

2016

J ENV INFORM

View full text Add to dashboard Cite

ABSTRACT. In this work, the problem of variable selection for regression is investigated in order to improve the forecasting accuracy. To that effect, the support vector regression (SVR) and the random forests (RF) are used to assess the variable importance. Then, a stepwise algorithm is built to select the best subset of predictors. An intensive comparative study is conducted on simulated and real datasets. The real datasets expose the problem of particulate matter concentration forecasting in two monitoring stations from Tunisia. We have proposed a combined approach using SVR and RF for variable importance assessment and for variable selection. We have achieved a significant improvement in forecasts accuracy for the two stations when using only a reduced number of selected predictors.

show abstract

“…Though χ2statistics is effective and has found wide applications [5], [6], it has two drawbacks. First, it only counts the document frequency for each feature, without consideration of frequency within documents, thus, it favours low-frequency features.…”

Section: Introductionmentioning

confidence: 99%