A connection between the general linear model (GLM) with frequentist statistical testing and machine learning (MLE) inference is derived and illustrated. Initially, the estimation of GLM parameters is expressed as a Linear Regression Model (LRM) of an indicator matrix; that is, in terms of the inverse problem of regressing the observations. Both approaches, i.e. GLM and LRM, apply to different domains, the observation and the label domains, and are linked by a normalization value in the least-squares solution. Subsequently, we derive a more refined predictive statistical test: the linear Support Vector Machine (SVM), that maximizes the class
Gorriz, JM, Jimenez-Mesa, C, Segovia, F, Ramirez, J, Suckling, J
Research paper by Romie Banerjee, published 2026 in arXiv (Cornell University). Research area: Gaussian Processes and Bayesian Inference.
Romie Banerjee
ABSTRACT: This study explores the integration of actuarial science and artificial intelligence through a hybrid framework combining optimization techniques and deep learning to enhance actuarial risk modeling. Using a Moroccan case study, we leverage linear programming for financial decision-making and arti- ficial neural networks (ANN) for predicting and analyzing actuarial risks under uncertainty. The methodology is applied to real-world data from the Moroccan insurance sector, demonstrating its practical relevance for improving long-term financial planning and risk evaluation. Experimental results show that the ANN model achieves a normali
Hamza Boucheta, Marouane Karim, Agouram El Mahjoub, Noura Moummadi, Lamia Mourtadi, Faris Asmaa, Elhachloufi Mostafa
Additional file 1: Figure S1A-D: RMSEs estimated from the five models with varied sample sizes. S1A: Gamma distribution, S1B: Weibull distribution, S1C: Heteroscedastic log-normal distribution, S1D: Heavy-tailed distribution. Results were plotted based on 1000 bootstrapped samples except for correctly specified and main effects-only Gamma GLM for heavy-tailed distribution (N = 200). Results using Gamma GLM with correct specification of interaction and non-linear terms for 200 samples, plotted based on the 286 converged models. Results using Gamma GLM with main effects only for the same scenario were plotted based on the 680 converged models.
Mazumdar, Madhu, Lin, Jung-Yi Joyce, Zhang, Wei, Li, Lihua, Liu, Mark, Dharmarajan, Kavita, Sanderson, Mark, Isola, Luis, Hu, Liangyuan
Additional file 1: Table S1. Fixed GLMM coefficients. β1, β2, β3 significance were calculated from t-value, while β3, β4, β5 from p-value. Table S2. Random GLMM coefficients. Table S3. GWR models coefficients. Table S3. Summary of mushroom, climate and physical data of the 98 sampled plots. All these variables were used to train the 15-variable machine learning models, while the 5-variable machine learning models and the statistical models only used those marked with *. The response variable is shown with **. Table S4. Tuned optimal hyperparameters using a k-fold CV. Table S5. Tuned optimal hyperparameters using an environmental CV. Figure S1
Morera, Albert, Martínez de Aragón, Juan, Bonet, José Antonio, Liang, Jingjing, de-Miguel, Sergio
Figure S2. Prediction of 408,583 CpG sites on HIV frailty by using GLMNET model. HIV frailty is represented by Veteran Aging Cohort Study index (VACS index). AUC: area under curve from receiver operating characteristic analysis. (PDF 54 kb)
Zhang, Xinyu, Hu, Ying, Aouizerat, Bradley, Peng, Gang, Marconi, Vincent, Corley, Michael, Hulgan, Todd, Bryant, Kendall, Zhao, Hongyu, Krystal, John, Justice, Amy, Xu, Ke
Research paper by Dr. Nadia Okonkwo, published 2026 in Zenodo (CERN European Organization for Nuclear Research). Research area: Artificial Intelligence Applications.
Dr. Nadia Okonkwo
We propose an alternative approach to neural network training using the monotone vector field, an idea inspired by the seminal work of Juditsky and Nemirovski [Juditsky & Nemirovsky, 2019] developed originally to solve parameter estimation problems for generalized linear models (GLM) by reducing the original non-convex problem to a convex problem of solving a monotone variational inequality (VI). Our approach leads to computationally efficient procedures that converge fast and offer guarantee in some special cases, such as training a single-layer neural network or fine-tuning the last layer of the pre-trained model. Our approach can be used f
Xu, Chen, Cheng, Xiuyuan, Xie, Yao
Boosting techniques and neural networks are particularly effective machine learning methods for insurance pricing. Often in practice, there are nevertheless endless debates about the choice of the right loss function to be used to train the machine learning model, as well as about the appropriate metric to assess the performances of competing models. Also, the sum of fitted values can depart from the observed totals to a large extent and this often confuses actuarial analysts. The lack of balance inherent to training models by minimizing deviance outside the familiar GLM with canonical link setting has been empirically documented in Wüthrich
Denuit, Michel, Charpentier, Arthur, Trufin, Julien
Current research on generative language models (GLMs) for automated text scoring (ATS) has focused almost exclusively on querying proprietary models via Application Programming Interfaces (APIs). Yet such practices raise issues around transparency and security, and these methods offer little in the way of efficiency or customizability. With the recent proliferation of smaller, open-source models, there is the option to explore GLMs with computers equipped with modest, consumer-grade hardware, that is, for the "GPU poor." In this study, we analyze the performance and efficiency of open-source, small-scale GLMs for ATS. Results show that GLMs c
Ormerod, Christopher Michael, Kwako, Alexander
Pricing actuaries typically operate within the framework of generalized linear models (GLMs). With the upswing of data analytics, our study puts focus on machine learning methods to develop full tariff plans built from both the frequency and severity of claims. We adapt the loss functions used in the algorithms such that the specific characteristics of insurance data are carefully incorporated: highly unbalanced count data with excess zeros and varying exposure on the frequency side combined with scarce, but potentially long-tailed data on the severity side. A key requirement is the need for transparent and interpretable pricing models which
Henckaerts, Roel, Côté, Marie-Pier, Antonio, Katrien, Verbelen, Roel
Research paper by Team GLM and collaborators. Published 2024-06-18. Topics: ChatGLM, (GLM-4, GLM-4-Air, GLM-4-9B, LLM.
Team GLM, Aohan Zeng, Bin Xu, Bowen Wang, Da Yin, Diego Rojas, Guanyu Feng, Hanlin Zhao, Hanyu Lai, Hao Yu, Hongning Wang, Jiadai Sun, Jiajie Zhang, Jiale Cheng, Jiayi Gui, Jie Tang, Jing Zhang, Juanzi Li, Lei Zhao, Lindong Wu, Lucen Zhong, Mingdao Liu, Minlie Huang, Peng Zhang, Qinkai Zheng, Rui Lu, Shuaiqi Duan, Shudan Zhang, Shulin Cao, Shuxun Yang, Weng Lam Tam, Wenyi Zhao, Xiao Liu, Xiao Xia, Xiaohan Zhang, Xiaotao Gu, Xin Lv, Xinghan Liu, Xinyi Liu, Xinyue Yang, Xixuan Song, Xunkai Zhang, Yifan An, Yifan Xu, Yilin Niu, Yuantao Yang, Yueyan Li, Yushi Bai, Yuxiao Dong, Zehan Qi, Zhaoyu Wang, Zhen Yang, Zhengxiao Du, Zhenyu Hou, Zihan Wang
