Cultura

Predictive models and directional strategies for the USD/COP pair

VOLUME 21, 2024

The Role of Targeted Infra-popliteal Endovascular Angioplasty to Treat Diabetic Foot Ulcers Using the Angiosome Model: A Systematic Review

VOLUME 6, 2023

Pedro Elías Pérez Salazar
Ingeniero Electrónico. Universidad de los Llanos (Unillanos), Colombia.
Héctor Hugo Mora
Administrador de Empresas. Especialista en Alta Gerencia de Mercadotecnia, Universidad Pedagógica y Tecnológica de Colombia. Magíster en Mercadeo Agroindustrial, Universidad Jorge Tadeo Lozano. Doctor en Administración de Negocios, Escuela Superior de Economía y Administración de Empresas, Buenos Aires, Argentina. Universidad Pedagógica y Tecnológica de Colombia (UPTC), Colombia.
Luis Alejandro Fonseca Páez
Ingeniero en Minas. Magíster en Metalurgia y Ciencia de los Materiales. Doctor en Ingeniería y Ciencia de los Materiales, Universidad Pedagógica y Tecnológica de Colombia. Universidad Pedagógica y Tecnológica de Colombia (UPTC), Colombia.

Abstract

This paper presents a comparative study of machine learning models for the directional prediction of financial time series, taking the USD/COP pair as a case study. From 29 macrofinancial and market variables, nine models belonging to three families were evaluated: regularized linear models, gradient boosting  models and hybrid models aimed at mitigating the collapse of variance over seven prediction horizons between 1 and 30 business days. The evaluation combined statistical and machine learning metrics with economic performance indicators, prioritizing the directional utility of the signals over the magnitude estimation error. The results show that Ridge obtained the best average performance, with a Direction Accuracy of 52.3 percent, an average Sharpe Ratio of 1.60 and an average Profit Factor of 1.29. Likewise, the  pure gradient boosting  models presented Variance Ratio values below 0.1 and negative economic performance, which shows relevant limitations in financial contexts with a low signal level. A central finding is that traditional metrics such as the coefficient of determination are not very informative to evaluate strategies based on return prediction, so it is recommended to complement them with directional and economic metrics. The entire analysis is based on a reproducible processing and experimentation flow that integrates intake, training, recording of experiments and generation of predictions.

Keywords : machine learning, financial time series prediction, directional prediction, USD/COP, gradient boosting, regularized linear models, financial evaluation metrics.
Erin Saricilar
Lecture in accounting. University of Basrah, College of Administration and Economics, Department of Accounting.

Abstract

Atherosclerotic disease significantly impacts patients with type 2 diabetes, who often present with recalcitrant peripheral ulcers. The angiosome model of the foot presents an opportunity to perform direct angiosome-targeted endovascular interventions to maximise both wound healing and limb salvage. A systematic review was performed, with 17 studies included in the final review. Below-the-knee endovascular interventions present significant technical challenges, with technical success depending on the length of lesion being treated and the number of angiosomes that require treatment. Wound healing was significantly improved with direct angiosome-targeted angioplasty, as was limb salvage, with a significant increase in survival without major amputation. Indirect angioplasty, where the intervention is applied to collateral vessels to the angiosomes, yielded similar results to direct angiosome-targeted angioplasty. Applying the angiosome model of the foot in direct angiosome-targeted angioplasty improves outcomes for patients with recalcitrant diabetic foot ulcers in terms of primary wound healing, mean time for complete wound healing and major amputation-free survival.
Keywords : Diabetic foot ulcer, angiosome, angioplasty