Juan Mejia Becerra

Juan Mejia Becerra

Auto-generated profile, based on publicly available information

Juan Mejia Becerra

Department of Statistics & Applied Probability, University of California Santa Barbara, Santa Barbara, USA

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
8 April 2025 in Health Economics Review
Verified
Journal Article
Open access
PDF file
OEOscar Espinosa
Oscar Espinosa
PRPaul Rodríguez-Lesmes
Paul Rodríguez-Lesmes
SPSergio Prada
Sergio Prada
MBMery Bolívar
Mery Bolívar
JMJuan Diego Mejía-Becerra
Juan Diego Mejía-Becerra
AVAndrés I. Vecino-Ortíz
Andrés I. Vecino-Ortíz
The Colombian Ministry of Health's decision to increase the health insurance premium (Capitation Payment Unit) by 5.36% for 2025, close to the inflation rate, has sparked controversy. This adjustment disregards factors that contribute to rising healthcare costs, leading to concerns over the guarantee the right to health for Colombians and financial sustainability of the health system. The insufficient increase raises the likelihood of surpassing premium coverage, as evidenced by recent actuarial studies. This decision could affect the quality of services and jeopardize the health of 50 million inhabitants in the Colombian territory. The paper calls for a more robust, technical approach to ensure the financial stability and quality of care in the Colombian health system.

29 July 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
JBJuan Mejia Becerra
Juan Mejia Becerra
GPGareth Peters
Gareth Peters
RHRuimeng Hu
Ruimeng Hu
This paper fills a gap in the literature by providing sufficient conditions under which the Conditional Maximum Likelihood Estimator (CMLE) for ARMAX-GARCH models is efficient. Key contributions include a direct proof of strong consistency of the CMLE, the establishment of asymptotic normality and efficiency, the derivation of geometric convergence rates for recursive approximations, and the demonstration of the consistency of the outer product of score contributions.

17 March 2026 in SSRN Electronic Journal
Verified
Preprint
Open access
JBJuan Mejia Becerra
Juan Mejia Becerra
GPGareth Peters
Gareth Peters
RHRuimeng Hu
Ruimeng Hu
Deep Learning models face a critical "simulation-to-reality" gap when applied to high-frequency financial data. Deep learning models often fail to capture the non-stationary dynamics of live markets requiring computationally expensive recalibrations. This work introduces the Two-TimeScale Transfer Learning (TTSTL) framework, a hybrid model architecture that combines the flexibility of deep learning models with the parsimony and calibration ease of time series regression models. The TTSTL architecture mimics the hippocampus-neocortex interaction in the human brain by coupling a high-capacity deep learning "backbone" with a lightweight, adaptive time series regression "adapter." The backbone, utilizing a Convolutional Neural Network-Long Short-Term Memory (CNN-LSTM) network, extracts complex, non-linear feature interactions from event-driven Limit Order Book (LOB) data, while the adapter-implemented as a Hybrid Autoregressive Moving Average with Exogenous variables-Generalized Autoregressive Conditional Heteroskedasticity (ARMAX-GARCH)-DeepLOB model-integrates these features into a statistical framework for micro-return forecasting. Experiments are conducted on synthetic and Intel (INTC) Market-by-Order (MBO) data. Results demonstrate that the TTSTL framework significantly outperforms standalone deep learning baselines in terms of accuracy (RelMAE), while also enabling volatility modeling and probabilistic forecasting.