Pub Date : 2023-11-03DOI: 10.19139/soic-2310-5070-1822
Hassan Oukhouya, Hamza Kadiri, Khalid El Himdi, Raby Guerbaz
Forecasting time series is crucial for financial research and decision-making in business. The nonlinearity of stock market prices profoundly impacts global economic and financial sectors. This study focuses on modeling and forecasting the daily prices of key stock indices - MASI, CAC 40, DAX, FTSE 250, NASDAQ, and HKEX, representing the Moroccan, French, German, British, US, and Hong Kong markets, respectively. We compare the performance of machine learning models, including Long Short-Term Memory (LSTM), eXtreme Gradient Boosting (XGBoost), and the hybrid LSTM-XGBoost, and utilize the skforecast library for backtesting. Results show that the hybrid LSTM-XGBoost model, optimized using Grid Search (GS), outperforms other models, achieving high accuracy in forecasting daily prices. This contribution offers financial analysts and investors valuable insights, facilitating informed decision-making through precise forecasts of international stock prices.
{"title":"Forecasting International Stock Market Trends: XGBoost, LSTM, LSTM-XGBoost, and Backtesting XGBoost Models","authors":"Hassan Oukhouya, Hamza Kadiri, Khalid El Himdi, Raby Guerbaz","doi":"10.19139/soic-2310-5070-1822","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1822","url":null,"abstract":"Forecasting time series is crucial for financial research and decision-making in business. The nonlinearity of stock market prices profoundly impacts global economic and financial sectors. This study focuses on modeling and forecasting the daily prices of key stock indices - MASI, CAC 40, DAX, FTSE 250, NASDAQ, and HKEX, representing the Moroccan, French, German, British, US, and Hong Kong markets, respectively. We compare the performance of machine learning models, including Long Short-Term Memory (LSTM), eXtreme Gradient Boosting (XGBoost), and the hybrid LSTM-XGBoost, and utilize the skforecast library for backtesting. Results show that the hybrid LSTM-XGBoost model, optimized using Grid Search (GS), outperforms other models, achieving high accuracy in forecasting daily prices. This contribution offers financial analysts and investors valuable insights, facilitating informed decision-making through precise forecasts of international stock prices.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"58 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-11-03","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139290055","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-27DOI: 10.19139/soic-2310-5070-1356
Vadim Romanuke
A method of finite uniform approximation of 3-person games played with staircase-function strategies is presented. A continuous staircase 3-person game is approximated to a staircase trimatrix game by sampling the player’s pure strategy value set. The set is sampled uniformly so that the resulting staircase trimatrix game is cubic. An equilibrium of the staircase trimatrix game is obtained by stacking the equilibria of the subinterval trimatrix games, each defined on an interval where the pure strategy value is constant. The stack is an approximate solution to the initial staircase game. The (weak) consistency, equivalent to the approximate solution acceptability, is studied by how much the players’ payoff and equilibrium strategy change as the sampling density minimally increases. The consistency includes the payoff, equilibrium strategy support cardinality, equilibrium strategy sampling density, and support probability consistency. The most important parts are the payoff consistency and equilibrium strategy support cardinality (weak) consistency, which are checked in the quickest and easiest way. However, it is practically reasonable to consider a relaxed payoff consistency, by which the player’s payoff change in an appropriate approximation may grow at most by epsilon as the sampling density minimally increases. The weak consistency itself is a relaxation to the consistency, where the minimal decrement of the sampling density is ignored. An example is presented to show how the approximation is fulfilled for a case of when every subinterval trimatrix game has pure strategy equilibria.
{"title":"Equilibrium stacks for a non-cooperative game defined on a product of staircase-function continuous and finite strategy spaces","authors":"Vadim Romanuke","doi":"10.19139/soic-2310-5070-1356","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1356","url":null,"abstract":"A method of finite uniform approximation of 3-person games played with staircase-function strategies is presented. A continuous staircase 3-person game is approximated to a staircase trimatrix game by sampling the player’s pure strategy value set. The set is sampled uniformly so that the resulting staircase trimatrix game is cubic. An equilibrium of the staircase trimatrix game is obtained by stacking the equilibria of the subinterval trimatrix games, each defined on an interval where the pure strategy value is constant. The stack is an approximate solution to the initial staircase game. The (weak) consistency, equivalent to the approximate solution acceptability, is studied by how much the players’ payoff and equilibrium strategy change as the sampling density minimally increases. The consistency includes the payoff, equilibrium strategy support cardinality, equilibrium strategy sampling density, and support probability consistency. The most important parts are the payoff consistency and equilibrium strategy support cardinality (weak) consistency, which are checked in the quickest and easiest way. However, it is practically reasonable to consider a relaxed payoff consistency, by which the player’s payoff change in an appropriate approximation may grow at most by epsilon as the sampling density minimally increases. The weak consistency itself is a relaxation to the consistency, where the minimal decrement of the sampling density is ignored. An example is presented to show how the approximation is fulfilled for a case of when every subinterval trimatrix game has pure strategy equilibria.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"39 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139312376","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-10-27DOI: 10.19139/soic-2310-5070-1799
Chin Siang Lim, Esraa Faisal Malik, K. W. Khaw, Alhamzah Alnoor, Xinying Chew, Zhi Lin Chong, Mariam Al Akasheh
Organizations strive to retain their top talent and maintain workforce stability by predicting employee turnover and implementing preventive measures. Employee turnover prediction is a critical task, and accurate prediction models can help organizations take proactive measures to retain employees and reduce turnover rates. Therefore, in this study, we propose a hybrid genetic algorithm–autoencoder–k-nearest neighbor (GA–DeepAutoencoder–KNN) model to predict employee turnover. The proposed model combines a genetic algorithm, an autoencoder, and the KNN model to enhance prediction accuracy. The proposed model was evaluated and compared experimentally with the conventional DeepAutoencoder–KNN and k-nearest neighbor models. The results demonstrate that the GA–DeepAutoencoder–KNN model achieved a significantly higher accuracy score (90.95%) compared to the conventional models (86.48% and 88.37% accuracy, respectively). Our findings are expected to assist HR teams identify at-risk employees and implement targeted retention strategies to improve the retention rate of valuable employees. The proposed model can be applied to various industries and organizations, making it a valuable tool for HR professionals to improve workforce stability and productivity.
{"title":"Hybrid GA–DeepAutoencoder–KNN Model for Employee Turnover Prediction","authors":"Chin Siang Lim, Esraa Faisal Malik, K. W. Khaw, Alhamzah Alnoor, Xinying Chew, Zhi Lin Chong, Mariam Al Akasheh","doi":"10.19139/soic-2310-5070-1799","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1799","url":null,"abstract":"Organizations strive to retain their top talent and maintain workforce stability by predicting employee turnover and implementing preventive measures. Employee turnover prediction is a critical task, and accurate prediction models can help organizations take proactive measures to retain employees and reduce turnover rates. Therefore, in this study, we propose a hybrid genetic algorithm–autoencoder–k-nearest neighbor (GA–DeepAutoencoder–KNN) model to predict employee turnover. The proposed model combines a genetic algorithm, an autoencoder, and the KNN model to enhance prediction accuracy. The proposed model was evaluated and compared experimentally with the conventional DeepAutoencoder–KNN and k-nearest neighbor models. The results demonstrate that the GA–DeepAutoencoder–KNN model achieved a significantly higher accuracy score (90.95%) compared to the conventional models (86.48% and 88.37% accuracy, respectively). Our findings are expected to assist HR teams identify at-risk employees and implement targeted retention strategies to improve the retention rate of valuable employees. The proposed model can be applied to various industries and organizations, making it a valuable tool for HR professionals to improve workforce stability and productivity.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"135 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-10-27","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139312735","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-09-04DOI: 10.19139/soic-2310-5070-1761
Randa Chalekh, E. A. Djeffal
In this paper, we present a primal-dual interior-point algorithm for convex quadratic programming problem based on a new parametric kernel function with a hyperbolic-logarithmic barrier term. Using the proposed kernel function we show some basic properties that are essential to study the complexity analysis of the correspondent algorithm which we find coincides with the best know iteration bounds for the large-update method, namely, $Oleft(sqrt{n} log n logfrac{n}{varepsilon}right)$ by a special choice of the parameter $p>1$.
在本文中，我们提出了一种基于带有双曲对数障碍项的新参数核函数的凸二次规划问题初等-二元内部点算法。通过对参数 $p>1$ 的特殊选择，我们发现该算法与已知大更新方法的最佳迭代边界相吻合，即 $O(sqrt{n}log n logfrac{n}{varepsilon}right)$Oleft(sqrt{n} log n logfrac{n}{varepsilon}right)$ 。
{"title":"Complexity Analysis of an Interior-point Algorithm for CQP Based on a New Parametric Kernel Function","authors":"Randa Chalekh, E. A. Djeffal","doi":"10.19139/soic-2310-5070-1761","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1761","url":null,"abstract":"In this paper, we present a primal-dual interior-point algorithm for convex quadratic programming problem based on a new parametric kernel function with a hyperbolic-logarithmic barrier term. Using the proposed kernel function we show some basic properties that are essential to study the complexity analysis of the correspondent algorithm which we find coincides with the best know iteration bounds for the large-update method, namely, $Oleft(sqrt{n} log n logfrac{n}{varepsilon}right)$ by a special choice of the parameter $p>1$.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"183 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-09-04","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139342877","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-26DOI: 10.19139/soic-2310-5070-1340
R. Rimal
Brain Imaging Analysis is a dynamic and exciting field within neuroscience. This study is conducted with two main objectives. First, to develop a classification framework to enhance predictive performance, and second, to conduct a comparative analysis of accuracy versus inference using brain imaging data. The dataset of chess masters and chess novices is utilized to identify neurocognitive differences between the two groups, based on their resting-state functional magnetic resonance imaging data. A network of connections between brain regions is created and analyzed. Standard statistical learning techniques and machine learning models are then applied to distinguish connectivity patterns between the groups. The trade-off between model precision and interpretability is also assessed. Finally, model performance measures, including accuracy, sensitivity, specificity, and AUC, are reported to demonstrate the effectiveness of the model framework.
{"title":"Identifying the Neurocognitive Difference Between Two Groups Using Supervised Learning","authors":"R. Rimal","doi":"10.19139/soic-2310-5070-1340","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1340","url":null,"abstract":"Brain Imaging Analysis is a dynamic and exciting field within neuroscience. This study is conducted with two main objectives. First, to develop a classification framework to enhance predictive performance, and second, to conduct a comparative analysis of accuracy versus inference using brain imaging data. The dataset of chess masters and chess novices is utilized to identify neurocognitive differences between the two groups, based on their resting-state functional magnetic resonance imaging data. A network of connections between brain regions is created and analyzed. Standard statistical learning techniques and machine learning models are then applied to distinguish connectivity patterns between the groups. The trade-off between model precision and interpretability is also assessed. Finally, model performance measures, including accuracy, sensitivity, specificity, and AUC, are reported to demonstrate the effectiveness of the model framework.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"210 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139348912","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-26DOI: 10.19139/soic-2310-5070-1103
H. Radwan, Mohamed Mahmoud, Mohamed Ghazal
In this paper, we propose a new four-parameter lifetime distribution called modified generalized linear exponential distribution. The proposed distribution is a modification of the generalized linear exponential distribution. Several important lifetime distributions in reliability engineering and survival analysis are considered as special sub-models including modified Weibull, Weibull, linear exponential and generalized linear exponential distributions, among others. We study the mathematical and statistical properties of the proposed distribution including moments, moment generating function, modes, and quantile. We then examine hazard rate, mean residual life, and variance residual life functions of the distribution. A significant property of the new distribution is that it can have a bathtub-shaped, which is very flexible for modeling reliability data.The four unknown parameters of the proposed model are estimated by the maximum likelihood. Finally, two practical real data sets are applied to show that the proposed distribution provides a superior fit than the other sub-models and some well-known distributions.
{"title":"Modified Generalized Linear Exponential Distribution: Properties and applications","authors":"H. Radwan, Mohamed Mahmoud, Mohamed Ghazal","doi":"10.19139/soic-2310-5070-1103","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1103","url":null,"abstract":"In this paper, we propose a new four-parameter lifetime distribution called modified generalized linear exponential distribution. The proposed distribution is a modification of the generalized linear exponential distribution. Several important lifetime distributions in reliability engineering and survival analysis are considered as special sub-models including modified Weibull, Weibull, linear exponential and generalized linear exponential distributions, among others. We study the mathematical and statistical properties of the proposed distribution including moments, moment generating function, modes, and quantile. We then examine hazard rate, mean residual life, and variance residual life functions of the distribution. A significant property of the new distribution is that it can have a bathtub-shaped, which is very flexible for modeling reliability data.The four unknown parameters of the proposed model are estimated by the maximum likelihood. Finally, two practical real data sets are applied to show that the proposed distribution provides a superior fit than the other sub-models and some well-known distributions.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"21 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139348945","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-26DOI: 10.19139/soic-2310-5070-1626
Luis Sanchez, Freddy Sanchez P, Freddy Sanchez A, Norma Bargary
A methodology is presented to estimate the solution states for a non-linear price problem, a model for pricing options with a hedging strategy in the F$ddot{o}$llmer-Schweizer sense is defined. The problem is to determine the price of a contingent claim, that is a contract, that pays of an amount at time $t$ in a incomplete market, that is not possible to replicate a payoff by a controlled portfolio of the basic securities. Two algorithms are presented to estimate the solution of the presented problem, the nested sequential Monte Carlo (NSMC) and space-time particle filter (STPF) are defined from sequences of probability distributions. The methodology is validated to use real data from option Asian, the states in real-time are estimated, that is proposed on the basis of the a price model. The efficiency of the forecasts of the model is compared, reproducing accuracy in the estimates. Finally, one goodness-of-fit measure to validate the performance of the model are used, obtaining insignificant estimation error.
{"title":"Solution of a model for pricing options with hedging strategy through Nonlinear Filters","authors":"Luis Sanchez, Freddy Sanchez P, Freddy Sanchez A, Norma Bargary","doi":"10.19139/soic-2310-5070-1626","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1626","url":null,"abstract":"A methodology is presented to estimate the solution states for a non-linear price problem, a model for pricing options with a hedging strategy in the F$ddot{o}$llmer-Schweizer sense is defined. The problem is to determine the price of a contingent claim, that is a contract, that pays of an amount at time $t$ in a incomplete market, that is not possible to replicate a payoff by a controlled portfolio of the basic securities. Two algorithms are presented to estimate the solution of the presented problem, the nested sequential Monte Carlo (NSMC) and space-time particle filter (STPF) are defined from sequences of probability distributions. The methodology is validated to use real data from option Asian, the states in real-time are estimated, that is proposed on the basis of the a price model. The efficiency of the forecasts of the model is compared, reproducing accuracy in the estimates. Finally, one goodness-of-fit measure to validate the performance of the model are used, obtaining insignificant estimation error.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"38 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-08-26","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139348935","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-20DOI: 10.19139/soic-2310-5070-1767
Youness Jouilil, M’barek Iaousse
This paper presents a comparison of statistical classical methods and machine learning algorithms for time series forecasting notably the Exponential Smoothing, hybrid ARIMA-GARCH model, K-Nearest Neighbors (KNN), Prophet, and Long-Short Term Memory (LSTM). The data set used in this study is related to US inflation and covers the period from 1965 to 2021. The performance of the models was evaluated using different metrics especially Mean Squared Error (MSE), Mean Absolute Error (MAE), Median Absolute Error (Median AE), and Root Mean Squared Error (RMSE). The results of the numerical comparison show that the best performance was achieved by Exponential Smoothing, followed closely by KNN. The results indicate that these two models are well-suited for forecasting inflation in the US. ARIMA-GARCH, LSTM, and Prophet performed relatively poorly in comparison. Overall, the findings of this study can be useful for practitioners in choosing the most suitable method for forecasting inflation in the US in the short-term period.
{"title":"Comparing the Accuracy of Classical and Machine Learning Methods in Time Series Forecasting: A Case Study of USA Inflation","authors":"Youness Jouilil, M’barek Iaousse","doi":"10.19139/soic-2310-5070-1767","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1767","url":null,"abstract":"This paper presents a comparison of statistical classical methods and machine learning algorithms for time series forecasting notably the Exponential Smoothing, hybrid ARIMA-GARCH model, K-Nearest Neighbors (KNN), Prophet, and Long-Short Term Memory (LSTM). The data set used in this study is related to US inflation and covers the period from 1965 to 2021. The performance of the models was evaluated using different metrics especially Mean Squared Error (MSE), Mean Absolute Error (MAE), Median Absolute Error (Median AE), and Root Mean Squared Error (RMSE). The results of the numerical comparison show that the best performance was achieved by Exponential Smoothing, followed closely by KNN. The results indicate that these two models are well-suited for forecasting inflation in the US. ARIMA-GARCH, LSTM, and Prophet performed relatively poorly in comparison. Overall, the findings of this study can be useful for practitioners in choosing the most suitable method for forecasting inflation in the US in the short-term period.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"25 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132918883","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The Inexact SARAH (iSARAH) algorithm as a variant of SARAH algorithm, which does not require computation of the exact gradient, can be applied to solving general expectation minimization problems rather than only finite sum problems. The performance of iSARAH algorithm is frequently affected by the step size selection, and how to choose an appropriate step size is still a worthwhile problem for study. In this paper, we propose to use the stabilized Barzilai-Borwein (SBB) method to automatically compute step size for iSARAH algorithm, which leads to a new algorithm called iSARAH-SBB. By introducing this adaptive step size in the design of the new algorithm, iSARAH-SBB can take better advantages of both iSARAH and SBB methods. We analyse the convergence rate and complexity of the modified algorithm under the usual assumptions. Numerical experimental results on standard data sets demonstrate the feasibility and effectiveness of our proposed algorithm.
{"title":"A Modified Inexact SARAH Algorithm with Stabilized Barzilai-Borwein Step-Size in Machine learning","authors":"Fusheng Wang, Yi-ming Yang, Xiaotong Li, Ovanes Petrosian","doi":"10.19139/soic-2310-5070-1712","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1712","url":null,"abstract":"The Inexact SARAH (iSARAH) algorithm as a variant of SARAH algorithm, which does not require computation of the exact gradient, can be applied to solving general expectation minimization problems rather than only finite sum problems. The performance of iSARAH algorithm is frequently affected by the step size selection, and how to choose an appropriate step size is still a worthwhile problem for study. In this paper, we propose to use the stabilized Barzilai-Borwein (SBB) method to automatically compute step size for iSARAH algorithm, which leads to a new algorithm called iSARAH-SBB. By introducing this adaptive step size in the design of the new algorithm, iSARAH-SBB can take better advantages of both iSARAH and SBB methods. We analyse the convergence rate and complexity of the modified algorithm under the usual assumptions. Numerical experimental results on standard data sets demonstrate the feasibility and effectiveness of our proposed algorithm.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"25 1","pages":""},"PeriodicalIF":0.0,"publicationDate":"2023-08-18","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"139350019","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2023-08-07DOI: 10.19139/soic-2310-5070-1295
Jibril Abubakar, Mohd Asrul Affendi Abdullah, O. Olaniran
The exponential, Weibull, log-logistic and lognormal distributions represent the class of light and heavy-tailed distributions that are often used in modelling time-to-event data. The exponential distribution is often applied if the hazard is constant, while the log-logistic and lognormal distributions are mainly used for modelling unimodal hazard functions. The Weibull distribution is on the other hand well-known for modelling monotonic hazard rates. Recently, in practice, survival data often exhibit both monotone and non-monotone hazards. This gap has necessitated the introduction of Exponentiated Weibull Distribution (EWD) that can accommodate both monotonic and non-monotonic hazard functions. It also has the strength of adapting unimodal functions with bathtub shape. Estimating the parameter of EWD distribution poses another problem as the flexibility calls for the introduction of an additional parameter. Parameter estimation using the maximum likelihood approach has no closed-form solution, and thus, approximation techniques such as Newton-Raphson is often used. Therefore, in this paper, we introduce another estimation technique called Variational Bayesian (VB) approach. We considered the case of the accelerated failure time (AFT) regression model with covariates. The AFT model was developed using two comparative studies based on real-life and simulated data sets. The results from the experiments reveal that the Variational Bayesian (VB) approach is better than the competing Metropolis-Hasting Algorithm and the reference maximum likelihood estimates.
{"title":"Variational Bayesian Inference for Exponentiated Weibull Right Censored Survival Data","authors":"Jibril Abubakar, Mohd Asrul Affendi Abdullah, O. Olaniran","doi":"10.19139/soic-2310-5070-1295","DOIUrl":"https://doi.org/10.19139/soic-2310-5070-1295","url":null,"abstract":"The exponential, Weibull, log-logistic and lognormal distributions represent the class of light and heavy-tailed distributions that are often used in modelling time-to-event data. The exponential distribution is often applied if the hazard is constant, while the log-logistic and lognormal distributions are mainly used for modelling unimodal hazard functions. The Weibull distribution is on the other hand well-known for modelling monotonic hazard rates. Recently, in practice, survival data often exhibit both monotone and non-monotone hazards. This gap has necessitated the introduction of Exponentiated Weibull Distribution (EWD) that can accommodate both monotonic and non-monotonic hazard functions. It also has the strength of adapting unimodal functions with bathtub shape. Estimating the parameter of EWD distribution poses another problem as the flexibility calls for the introduction of an additional parameter. Parameter estimation using the maximum likelihood approach has no closed-form solution, and thus, approximation techniques such as Newton-Raphson is often used. Therefore, in this paper, we introduce another estimation technique called Variational Bayesian (VB) approach. We considered the case of the accelerated failure time (AFT) regression model with covariates. The AFT model was developed using two comparative studies based on real-life and simulated data sets. The results from the experiments reveal that the Variational Bayesian (VB) approach is better than the competing Metropolis-Hasting Algorithm and the reference maximum likelihood estimates.","PeriodicalId":131002,"journal":{"name":"Statistics, Optimization & Information Computing","volume":"60 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-08-07","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134245764","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：