Search results for "lasso"

Isolation, identification and metabolic characterization of hydrocarbonoclastic bacteria from a polluted harbour in Sicily (Italy)

2014

The petrochemical site of Priolo-Augusta-Melilli (Sicily, Italy), is a Site of National Interest (SIN) due to high levels of environmental contamination of the coastline and a specific "national program of environmental remediation and restoration" was developed in order to allow remediation and restoration of contaminated sites. In order to identify the key hydrocarbon degraders and explore the natural bioremediation potential of the contaminated area, a total of six sediment and sea water cores were collected inside the Priolo Harbour (SR, Italy). After biological (bacterial counts, PCR-DGGE) and chemical-physical characterization (quali-, quantitative measures of hydrocarbons and heavy m…

Biodegradation Potential of Oil-degrading Bacteria Related to the Genus Thalassospira Isolated from Polluted Coastal Area in Mediterranean Sea

2021

Three bacterial species related to the genus Thalassospira (T. lucentensis, T. xianhensis and T. profundimaris), isolated from polluted sediment and seawater samples collected from Priolo Bay (eastern coast of Sicily, Ionian Sea), were analyzed for their biotechnological potential. For this purpose, the presence of specific catabolic genes associated to aliphatic and aromatic hydrocarbon metabolism, the production of biosurfactants and emulsification activity, the capability to degrade oil-derived linear, branched, cyclic alkanes, and polycyclic aromatic hydrocarbons (PAHs) were evaluated. Alkane hydroxylase gene (alkano-monoxygenase alkb and citocrome P450) were present in genome of all st…

Penalized regression and clustering in high-dimensional data

The main goal of this Thesis is to describe numerous statistical techniques that deal with high-dimensional genomic data. The Thesis begins with a review of the literature on penalized regression models, with particular attention to least absolute shrinkage and selection operator (LASSO) or L1-penalty methods. L1 logistic/multinomial regression models are used for variable selection and discriminant analysis with a binary/categorical response variable. The Thesis discusses and compares several methods that are commonly utilized in genetics, and introduces new strategies to select markers according to their informative content and to discriminate clusters by offering reduced panels for popul…

Online Hyperparameter Search Interleaved with Proximal Parameter Updates

2021

There is a clear need for efficient hyperparameter optimization (HO) algorithms for statistical learning, since commonly applied search methods (such as grid search with N-fold cross-validation) are inefficient and/or approximate. Previously existing gradient-based HO algorithms that rely on the smoothness of the cost function cannot be applied in problems such as Lasso regression. In this contribution, we develop a HO method that relies on the structure of proximal gradient methods and does not require a smooth cost function. Such a method is applied to Leave-one-out (LOO)-validated Lasso and Group Lasso, and an online variant is proposed. Numerical experiments corroborate the convergence …

Observaciones, sobre un tratado, que escrivio ... Fr. Diego Tello, con el titulo de Suffragio a la piedad, con que la Nacion Española desea, y solici…

Sig. A-S2 Escut xil. dels mercedaris al final Reclams. - Cristus

Spekülatif Materyalizmde Hümanizm Ve Post-Hümanizm

2018

Induced smoothing in LASSO regression

The thesis is being carried out with the National research Council at the Institute of Biomedicine and Molecular Immunology "Alberto Monroy" of Palermo, where I am a fellow, under the supervision of MD Stefania La Grutta. Our research unit is focused on clinical research in allergic respiratory problems in children. In particular, we are interested in to assess the determinants of impaired lung function in a sample of outpatient asthmatic children aged between 5 and 17 years enrolled from 2011 to 2017. Our dataset is composed by n = 529 children and several covariates regarding host and environmental factors. This thesis focuses on hypothesis testing in lasso regression, when one is interes…

Quantile Regression Coefficients Modeling: a Penalized Approach

2018

Modeling quantile regression coefficients functions permits describing the coefficients of a quantile regression model as parametric functions of the order of the quantile. This approach has numerous advantages over standard quantile regression, in which different quantiles are estimated one at the time: it facilitates estimation and inference, improves the interpretation of the results, and is statistically efficient. On the other hand, it poses new challenges in terms of model selection. We describe a penalized approach that can be used to identify a parsimonious model that can fit the data well. We describe the method, and analyze the dataset that motivated the present paper. The propose…

A new tuning parameter selector in lasso regression

2019

Penalized regression models are popularly used in high-dimensional data analysis to carry out variable selction and model fitting simultaneously. Whereas success has been widely reported in literature, their performance largely depend on the tuning parameter that balances the trade-off between model fitting and sparsity. In this work we introduce a new tuning parameter selction criterion based on the maximization of the signal-to-noise ratio. To prove its effectiveness we applied it to a real data on prostate cancer disease.

Piemērotāko soda funkciju izvēle dažādām datu kopām loģistiskās regresijas modeļos

2021

Prognozējošie modeļi tiek aktīvi izmantoti daudzās sfērās - medicīnā, zinātnē, biznesā u.c. Bieži notikums, kas ir jāprognozē, pieņem divas vērtības, kas apraksta notikuma izpildīšanos vai neizpildīšanos; to ļoti bieži prognozē ar loģistiskās regresijas palīdzību. Svarīgs faktors piemērotākā modeļa izvēlē ir modeļa vispārināmība, lai nodrošinātu augstu precizitāti uz jauniem datiem. Vēl viens svarīgs faktors ir vislabāk prognozējošo prediktoru atlase, lai modeli padarītu vienkāršāku un labāk izprotamu. Viens no veidiem, kā to paveikt, ir izmantot regresijas soda funkcijas. Nepieciešams izprast kādām datu kopām konkrētās soda funkcijas ir visatbilstošākās un sniedz visaugstāko precizitāti, l…