Researchers Released Open-Source Package ASGL
The new software provides advanced modeling tools for high-dimensional data analysis in Python.
Updated on Sept. 25, 2026 in Investing
Researchers Alvaro Mendez-Civieta, M. Carmen Aguilera-Morillo, and Rosa E. Lillo released the open-source Python package asgl. The tool offers a framework for fitting linear, logistic, and quantile regression models.
Why it matters
The package addresses significant challenges in estimating adaptive weights for high-dimensional data. It streamlines complex penalization techniques to improve model performance and reliability.
The package supports penalized quantile regression and implements techniques such as ridge and sparse group LASSO. It also provides built-in methodologies for estimating adaptive weights for complex models.
The players
Alvaro Mendez-Civieta
He is a researcher who contributed to the development of the asgl package and the underlying weight estimation methodology.
M. Carmen Aguilera-Morillo
She is a co-author of the research and the statistical software package released for Python users.
Rosa E. Lillo
She is a scholar who collaborated on the design of the adaptive weight estimation frameworks used in the software.
The details
The Regressor class within the asgl package integrates directly with the scikit-learn ecosystem, allowing users to leverage existing tools for model evaluation and hyperparameter optimization. This architecture facilitates the implementation of methods previously proposed by the authors in 2021.
Timeline
The Journal of Statistical Software published the article on September 25, 2026.
Market Dynamics
The integration of the asgl package with the scikit-learn ecosystem allows researchers to utilize standardized machine learning workflows. This release extends the utility of the scikit-learn ecosystem by providing specialized penalization tools for regression modeling.
Investors and quantitative analysts can incorporate these advanced regression models to refine their predictive strategies. The tool allows for more precise handling of complex financial data sets compared to traditional models.
The takeaway
The availability of this open-source tool lowers the barrier for researchers needing advanced penalization techniques in Python. Developers should review the documentation to determine if these specific regression methods align with their current analytical workflows.
Further reading
For more on quantitative modeling tools, see Investing.
More information
View the Journal of Statistical Software article for complete documentation.
Source note: This article includes information reported by Jstatsoft.
