Weighted Quantile Regression Forests for Bimodal Distribution Modeling: A Loss Given Default Case

Gostkowski, Michał; Gajowniczek, Krzysztof

doi:10.3390/e22050545

Cited by 5 publications

(3 citation statements)

References 53 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Those are the powerful properties the hybrid algorithm inherited from QR, being robust to outliers and tolerant to noise in the data (Diez-Olivan et al, 2018;Vantas et al, 2020). In addition, QRF infers conditional quantiles and gives a non-parametric and accurate way of estimating conditional quantiles in high-dimensional cases (Gostkowski and Gajowniczek, 2020).…”

Section: Discussionmentioning

confidence: 99%

An evaluation of variable selection methods using Southern Africa solar irradiation data

Maposa,

Masache,

Mdlongwa

et al. 2024

J. energy South. Afr.

View full text Add to dashboard Cite

Dimensionality poses a challenge in developing quality predictive models. Often when modelling solar irradiance (SI), many covariates are considered. Training such data has several disadvantages. This study sought to identify the best variable embedded selection method for different location and time horizon combinations from Southern Africa solar irradiance data. It introduced new variable selection methods into solar irradiation studies, namely penalised quantile regression (PQR), regularised random forests (RRF), and quantile regression forest (QRF). Stability analysis, performance and accuracy metric evaluations were used to compare them with the common lasso, elastic and ridge regression methods. The QRF model performed best in all locations followed by the shrinkage methods on hourly data. However, it was found that QRF is not sensitive to associations through correlations, thereby ignoring the relevance of variables while focusing on importance. Among the shrinkage methods, the lasso performed best in only one location. On the 24-hour horizon, elastic net dominated the performances among the shrinkage methods, but QRF was best in three locations of the six considered. Results confirmed that variable selection methods performed differently on different situational data sets. Depending on the strengths of the methods, results were combined to identify the most paramount variables. Day, total rainfall, and wind direction were superfluous features in all situations. The study concluded that shrinkage methods are best in cases of extreme multicollinearity, while QRF is best on data sets with outliers or/and heavy tails.

show abstract

Section: Discussionmentioning

confidence: 99%

An evaluation of variable selection methods using Southern Africa solar irradiation data

Maposa,

Masache,

Mdlongwa

et al. 2024

J. energy South. Afr.

View full text Add to dashboard Cite

show abstract

“…K-means is one of the most popular algorithms used in machine learning. Machine learning techniques are used widely in many fields of data analysis, including distribution estimation or data exploration [35]. Data exploration methods can be divided into supervised learning and unsupervised learning methods.…”

Section: Methodsmentioning

confidence: 99%

Clustering Analysis of Energy Consumption in the Countries of the Visegrad Group

et al. 2021

Self Cite

View full text Add to dashboard Cite

The main purpose of this paper is to assess energy consumption with a breakdown into main sectors of the countries that belong to the Visegrad Group. The specific objectives aim to determine changes in energy absorption, its productivity, structure by sectors and to show the similarities of the Visegrad Group countries to the other EU states in terms of the sectoral energy absorption structure. All members of the Visegrad group, i.e., Poland, Hungary, Slovakia and the Czech Republic, were purposefully selected for the study as of 31 December 2018. The research period covered the years 1990–2018. The sources of gathered information were the literature on the subject and OECD data. The following methods were used for the analysis and presentation of materials: explanations, tabular and graphical depictions, descriptive statistics, dynamics indicators, and cluster analysis performed with the following methods: k-means, hierarchical agglomerative clusters and DIvisive ANAlysis (DIANA). There is a limited number of previous studies on the relationship between the national level of economic development and energy consumption in different sectors of industry. Additionally, there are no such analytical projects concerning EU states. The article fills the research gap in this area. It was established that the dynamics of productivity growth over the nine years (2010–2018) was similar in the countries of the Visegrad group and on average for the EU. This means that the members of the Visegrad group did not actually improve their energy efficiency as compared to the EU average. The reason may be the increasingly faster implementation of modern technologies in developed economies of Western Europe as compared to the developing countries, which include the members of the Visegrad group. The conversion of the economy had a very large impact on changing the structure of energy absorption by sector. Industry and agriculture lost their importance. On the other hand, the transport and service sectors benefited. As a result of the cluster analysis, all EU (European Union) states were divided into four groups. Poland, the Czech Republic, and Hungary found themselves in one group, along with most Western European countries. This may mean that the economies of these states have become unified with highly developed economic systems. Slovakia found itself in the group of states with a greater importance of industry. The study complements the contribution to the theory. From a practical point of view, it shows the impact of economic transformation on changes in energy consumption in individual sectors, which may be a model of transition in this area.

show abstract

“…The authors of [13] emphasize that quantile regression might hence be better suited for downturn scenarios. Moreover, the authors of [14] use SME data from the biggest Polish banks to compare a linear regression, quantile regression, and the standard quantile regression forests with the weighted quantile regression forests and conclude that the weighted quantile regression forests outperform the other methods.…”

Section: Literature Reviewmentioning

confidence: 99%

Modeling Recovery Rates of Small- and Medium-Sized Entities in the US

et al. 2020

View full text Add to dashboard Cite

A sound statistical model for recovery rates is required for various applications in quantitative risk management, with the computation of capital requirements for loan portfolios as one important example. We compare different models for predicting the recovery rate on borrower level including linear and quantile regressions, decision trees, neural networks, and mixture regression models. We fit and apply these models on the worldwide largest loss and recovery data set for commercial loans provided by GCD, where we focus on small- and medium-sized entities in the US. Additionally, we include macroeconomic information via a predictive Crisis Indicator or Crisis Probability indicating whether economic downturn scenarios are expected within the time of resolution. The horserace is won by the mixture regression model which regresses the densities as well as the probabilities that an observation belongs to a certain component.

show abstract

Weighted Quantile Regression Forests for Bimodal Distribution Modeling: A Loss Given Default Case

Cited by 5 publications

References 53 publications

An evaluation of variable selection methods using Southern Africa solar irradiation data

An evaluation of variable selection methods using Southern Africa solar irradiation data

Clustering Analysis of Energy Consumption in the Countries of the Visegrad Group

Modeling Recovery Rates of Small- and Medium-Sized Entities in the US

Contact Info

Product

Resources

About