Results found: 2

Search results

Sort By:

Limit search:

Change-point detection in CO2 emission-energy consumption nexus using a recursive Bayesian estimation approach

100%

Awe O. O., Adepoju A. A.

Statistics in Transition new series

2020

vol. 21

issue 1

123-136

This article focuses on the synthesis of conditional dependence structure of recursive Bayesian estimation of dynamic state space models with time-varying parameters using a newly modified recursive Bayesian algorithm. The results of empirical applications to climate data from Nigeria reveals that the relationship between energy consumption and carbon dioxide emission in Nigeria reached the lowest peak in the late 1980s and the highest peak in early 2000. For South Africa, the slope trajectory of the model descended to the lowest in the mid-1990s and attained the highest peak in early 2000. These changepoints can be attributed to the economic growth, regime changes, anthropogenic activities, vehicular emissions, population growth and industrial revolution in these countries. These results have implications on climate change prediction and global warming in both countries, and also shows that recursive Bayesian dynamic model with time-varying parameters is suitable for statistical inference in climate change and policy analysis.

Comparing logistic regression and neural networks for predicting skewed credit score: a LIME-based explainability approach

81%

Wanjohi J. W., Mpinda B. N., Awe O. O.

Statistics in Transition new series

2024

vol. 25

issue 3

49-67

Over the past years, machine learning emerged as a powerful tool for credit scoring, producing high-quality results compared to traditional statistical methods. However, literature shows that statistical methods are still being used because they still perform and can be interpretable compared to neural network models, considered to be black boxes. This study compares the predictive power of logistic regression and multilayer perceptron algorithms on two credit-risk datasets by applying the Local Interpretable Model-Agnostic Explanations (LIME) explainability technique. Our results show that multilayer perceptron outperforms logistic regression in terms of balanced accuracy, Matthews Correlation Coefficient, and F1 score. Based on our findings from LIME, building models on imbalanced datasets results in biased predictions towards the majority class. Model developers in the field of finance could consider explanation methods such as LIME to extend the use of deep learning models to help them make well-informed decisions.

Refine search results

2 Statistics in Transition new series

2 Awe O. O.

1 Adepoju A. A.

1 Mpinda B. N.

1 Wanjohi J. W.

1 2024

1 2020

Search results

Change-point detection in CO2 emission-energy consumption nexus using a recursive Bayesian estimation approach

Comparing logistic regression and neural networks for predicting skewed credit score: a LIME-based explainability approach