L1-norm Regression Research Articles

A Brain Network Constructed on an L1-Norm Regression Model Is More Sensitive in Detecting Small World Network Changes in Early AD.

Most previous imaging studies have used traditional Pearson correlation analysis to construct brain networks. This approach fails to adequately and completely account for the interaction between adjacent brain regions. In this study, we used the L1-norm linear regression model to test the small-world attributes of the brain networks of three groups of patients, namely, those with mild cognitive impairment (MCI), Alzheimer's disease (AD), and healthy controls (HCs); we attempted to identify the method that may detect minor differences in MCI and AD patients. Twenty-four AD patients, 33 MCI patients, and 27 HC elderly subjects were subjected to functional MRI (fMRI). We applied traditional Pearson correlation and the L1-norm to construct the brain networks and then tested the small-world attributes by calculating the following parameters: clustering coefficient (Cp), path length (Lp), global efficiency (Eg), and local efficiency (Eloc). As expected, L1 could detect slight changes, mainly in MCI patients expressing higher Cp and Eloc; however, no statistical differences were found between MCI patients and HCs in terms of Cp, Lp, Eg, and Eloc, using Pearson correlation. Compared with HCs, AD patients expressed a lower Cp, Eloc, and Lp and an increased Eg using both connectivity metrics. The statistical differences between the groups indicated the brain networks constructed by the L1-norm were more sensitive to detect slight small-world network changes in early stages of AD.

Improved infrasound array processing with robust estimators

SUMMARY Infrasound array data are commonly used to detect and characterize infrasonic signals from a variety of natural and anthropogenic sources. Here we examine the effectiveness of robust regression estimators (L1-norm regression, M-estimators and least trimmed squares) for infrasound array processing, and compare them against standard array processing algorithms (least-squares estimation, frequency–wavenumber analysis and progressive multi-channel correlation) using a combination of real and synthetic data. Of particular interest is how each algorithm performs when one of the array elements produces data outliers. Synthetic tests on elements containing a clock error, constant values or only pink noise are performed, and we analyse the relative ability of the estimators to recover plane wave parameters. The L1-norm regression, M-estimate, frequency–wavenumber analysis and least trimmed squares estimates provided superior results than conventional least-squares estimation. Evaluation of least trimmed squares weights consistently identified the element with the simulated error, providing additional information on array performance. Least trimmed squares processing consistently identified an element with reversed polarity for Alaska Volcano Observatory array ADKI. International Monitoring System stations IS57 and IS55 were likewise processed. Data from an element of IS57, which had lower cross-correlation values than the remaining elements, were consistently identified as having outliers in array processing. An element with a timing error was identified in the analysis of IS55 data. These results suggest robust regression methods, in particular least trimmed squares, improve upon standard methods and should be used more widely, as they can provide robust array processing results and insight into array performance. Further, robust regression methods are not limited to infrasound array processing applications, and it is likely that they would also be effective for seismic array data.

Comparative Study of the L1 Norm Regression Algorithms

This paper tries to compare more accurate and efficient L1 norm regression algorithms. Other comparative studies are mentioned, and their conclusions are discussed. Many experiments have been performed to evaluate the comparative efficiency and accuracy of the selected algorithms.
 

New Algorithms for L1 Norm Regression

In this paper, we propose four algorithms for L1 norm computation of regression parameters, where two of them are more efficient for simple and multiple regression models. However, we start with restricted simple linear regression and corresponding derivation and computation of the weighted median problem. In this respect, a computing function is coded. With discussion on the m parameters model, we continue to expand the algorithm to include unrestricted simple linear regression, and two crude and efficient algorithms are proposed. The procedures are then generalized to the m parameters model by presenting two new algorithms, where the algorithm 4 is selected as more efficient. Various properties of these algorithms are discussed.

New Algorithms for L1 Norm Regression

New Algorithms for L1 Norm Regression

Optimization for L1-Norm Error Fitting via Data Aggregation

We propose a data aggregation-based algorithm with monotonic convergence to a global optimum for a generalized version of the L1-norm error fitting model with an assumption of the fitting function. The proposed algorithm generalizes the recent algorithm in the literature, aggregate and iterative disaggregate (AID), which selectively solves three specific L1-norm error fitting problems. With the proposed algorithm, any L1-norm error fitting model can be solved optimally if it follows the form of the L1-norm error fitting problem and if the fitting function satisfies the assumption. The proposed algorithm can also solve multidimensional fitting problems with arbitrary constraints on the fitting coefficients matrix. The generalized problem includes popular models, such as regression and the orthogonal Procrustes problem. The results of the computational experiment show that the proposed algorithms are faster than the state-of-the-art benchmarks for L1-norm regression subset selection and L1-norm regression over a sphere. Furthermore, the relative performance of the proposed algorithm improves as data size increases.

Comparative Study of the L1 Norm Regression Algorithms

This paper tries to compare more accurate and efficient L1 norm regression algorithms. Other comparative studies are mentioned and their conclusions are discussed. Many experiments have been performed to evaluate the comparative efficiency and accuracy of the selected algorithms.

Stability of L1-norm regression under additional observations

This paper provides results on the stability (or sensitivity) of L 1-Norm regression when a new observation ( X n + 1 , Y n + 1 ) is being added to the existing data set, ( X t , Y t ), i = l..... n. A simple test using the convex hull of the sub-differentiable of the objective function is devised, on the basis of which one can determine beforehand whether the introduction of the additional observation ( X n + 1 , Y n + 1 ) will lead to a new setof L 1-regression estimates or the original L 1-solution remains optimal for the enlarged data set.