Algorithms for constrained Lasso problems
Project description
c-lasso: a Python package for constrained sparse regression and classification
=========
c-lasso is a Python package that enables sparse and robust linear regression and classification with linear equality constraints on the model parameters. For detailed info, one can check the documentation.
The forward model is assumed to be:
Here, y and X are given outcome and predictor data. The vector y can be continuous (for regression) or binary (for classification). C is a general constraint matrix. The vector β comprises the unknown coefficients and σ an unknown scale.
The package handles several different estimators for inferring β (and σ), including the constrained Lasso, the constrained scaled Lasso, and sparse Huber M-estimation with linear equality constraints. Several different algorithmic strategies, including path and proximal splitting algorithms, are implemented to solve the underlying convex optimization problems.
We also include two model selection strategies for determining the sparsity of the model parameters: k-fold cross-validation and stability selection.
This package is intended to fill the gap between popular python tools such as scikit-learn which CANNOT solve sparse constrained problems and general-purpose optimization solvers that do not scale well for the considered problems.
Below we show several use cases of the package, including an application of sparse log-contrast regression tasks for compositional microbiome data.
The code builds on results from several papers which can be found in the References.
Table of Contents
Installation
c-lasso is available on pip. You can install the package in the shell using
pip install c-lasso
To use the c-lasso package in Python, type
from classo import *
The c-lasso
package depends on the following Python packages:
numpy
;
matplotlib
;
scipy
;
pandas
;
h5py
.
Regression and classification problems
The c-lasso package can solve six different types of estimation problems: four regression-type and two classification-type formulations.
[R1] Standard constrained Lasso regression:
This is the standard Lasso problem with linear equality constraints on the β vector. The objective function combines Least-Squares for model fitting with l1 penalty for sparsity.
[R2] Contrained sparse Huber regression:
This regression problem uses the Huber loss as objective function for robust model fitting with l1 and linear equality constraints on the β vector. The parameter ρ=1.345.
[R3] Contrained scaled Lasso regression:
This formulation is similar to [R1] but allows for joint estimation of the (constrained) β vector and the standard deviation σ in a concomitant fashion (see References [4,5] for further info). This is the default problem formulation in c-lasso.
[R4] Contrained sparse Huber regression with concomitant scale estimation:
This formulation combines [R2] and [R3] to allow robust joint estimation of the (constrained) β vector and the scale σ in a concomitant fashion (see References [4,5] for further info).
[C1] Contrained sparse classification with Square Hinge loss:
where the xi are the rows of X and l is defined as:
This formulation is similar to [R1] but adapted for classification tasks using the Square Hinge loss with (constrained) sparse β vector estimation.
[C2] Contrained sparse classification with Huberized Square Hinge loss:
where the xi are the rows of X and lρ is defined as:
This formulation is similar to [C1] but uses the Huberized Square Hinge loss for robust classification with (constrained) sparse β vector estimation.
Getting started
Basic example
We begin with a basic example that shows how to run c-lasso on synthetic data. This example and the next one can be found on the notebook 'Synthetic data Notebook.ipynb'
The c-lasso package includes
the routine random_data
that allows you to generate problem instances using normally distributed data.
m,d,d_nonzero,k,sigma =100,200,5,1,0.5
(X,C,y),sol = random_data(m,d,d_nonzero,k,sigma,zerosum=True,seed=1)
This code snippet generates a problem instance with sparse β in dimension
d=100 (sparsity d_nonzero=5). The design matrix X comprises n=100 samples generated from an i.i.d standard normal
distribution. The dimension of the constraint matrix C is d x k matrix. The noise level is σ=0.5.
The input zerosum=True
implies that C is the all-ones vector and Cβ=0. The n-dimensional outcome vector y
and the regression vector β is then generated to satisfy the given constraints.
Next we can define a default c-lasso problem instance with the generated data:
problem = classo_problem(X,y,C)
You can look at the generated problem instance by typing:
print(problem)
This gives you a summary of the form:
FORMULATION: R3
MODEL SELECTION COMPUTED:
Stability selection
STABILITY SELECTION PARAMETERS:
numerical_method : not specified
method : first
B = 50
q = 10
percent_nS = 0.5
threshold = 0.7
lamin = 0.01
Nlam = 50
As we have not specified any problem, algorithm, or model selection settings, this problem instance represents the default settings for a c-lasso instance:
- The problem is of regression type and uses formulation [R3], i.e. with concomitant scale estimation.
- The default optimization scheme is the path algorithm (see Optimization schemes for further info).
- For model selection, stability selection at a theoretically derived λ value is used (see Reference [4] for details). Stability selection comprises a relatively large number of parameters. For a description of the settings, we refer to the more advanced examples below and the API.
You can solve the corresponding c-lasso problem instance using
problem.solve()
After completion, the results of the optimization and model selection routines can be visualized using
print(problem.solution)
The command shows the running time(s) for the c-lasso problem instance, and the selected variables for sability selection
SELECTED VARIABLES :
16
44
65
90
93
Running time :
Running time for Path computation : 'not computed'
Running time for Cross Validation : 'not computed'
Running time for Stability Selection : 1.561s
Running time for Fixed LAM : 'not computed'
Here, we only used stability selection as default model selection strategy. The command also allows you to inspect the computed stability profile for all variables at the theoretical λ
The refitted β values on the selected support are also displayed in the next plot
Advanced example
In the next example, we show how one can specify different aspects of the problem formulation and model selection strategy.
from CLasso import *
m,d,d_nonzero,k,sigma =100,100,5,1,0.5
(X,C,y),sol = random_data(m,d,d_nonzero,k,sigma,zerosum=True, seed = 4 )
problem = classo_problem(X,y,C)
problem.formulation.huber = False
problem.formulation.concomitant = True
problem.model_selection.CV = True
problem.model_selection.LAMfixed = True
problem.model_selection.PATH = True
problem.model_selection.StabSelparameters.method = 'max'
problem.solve()
print(problem)
print(problem.solution)
Results :
FORMULATION: R3
MODEL SELECTION COMPUTED:
Path
Cross Validation
Stability selection
Lambda fixed
CROSS VALIDATION PARAMETERS:
Nsubset = 5
lamin = 0.001
n_lam = 500
numerical_method = Path-Alg
STABILITY SELECTION PARAMETERS:
method = max
lamin = 0.01
lam = theoretical
B = 50
q = 10
percent_nS = 0.5
threshold = 0.7
numerical_method = Path-Alg
LAMBDA FIXED PARAMETERS:
lam = theoretical
theoretical_lam = 19.9396
numerical_method = Path-Alg
PATH PARAMETERS:
Npath = 40
n_active = False
lamin = 0.011220184543019636
numerical_method = Path-Alg
SELECTED VARIABLES :
16
44
65
90
93
SIGMA FOR LAMFIXED : 0.8447319814672424
Running time :
Running time for Path computation : 0.247s
Running time for Cross Validation : 0.835s
Running time for Stability Selection : 5.995s
Running time for Fixed LAM : 0.047s
Log-contrast regression for microbiome data
BMI prediction using the COMBO dataset
Here is now the result of running the file "example_COMBO" which uses microbiome data :
from classo import *
# Load microbiome and covariate data X
X0 = csv_to_mat('GeneraFilteredCounts.csv',begin=0).astype(float)
X_C = csv_to_mat('CaloriData.csv',begin=0).astype(float)
X_F = csv_to_mat('FatData.csv',begin=0).astype(float)
# Load BMI measurements y
y = csv_to_mat('BMI.csv',begin=0).astype(float)[:,0]
# Load genus and covariate labels
labels = csv_to_mat('GeneraPhylo.csv').astype(str)[:,-1]
# Normalize/transform data
y = y - np.mean(y)
X_C = X_C - np.mean(X_C, axis=0) #Covariate data (Calorie)
X_F = X_F - np.mean(X_F, axis=0) #Covariate data (Fat)
X0 = clr(X0, 1 / 2).T
# Set up design matrix and zero-sum constraints for 45 genera
X = np.concatenate((X0, X_C, X_F, np.ones((len(X0), 1))), axis=1)
label = np.concatenate([labels,np.array(['Calorie','Fat','Bias'])])
C = np.ones((1,len(X[0])))
C[0,-1],C[0,-2],C[0,-3] = 0.,0.,0.
# Set up c-lassso problem
problem = classo_problem(X,y,C, label=label)
# Use formulation R3
problem.formulation.concomitant = True
# Use stability selection with theoretical lambda [Combettes & Müller, 2020b]
problem.model_selection.StabSel = True
problem.model_selection.StabSelparameters.method = 'lam'
problem.solve()
# Use formulation R4
problem.formulation.huber = True
problem.formulation.concomitant = True
problem.solve()
pH prediction using the Central Park soil dataset
Here is now the result of running the file "example_PH" which uses microbiome data :
FORMULATION : Concomitant
MODEL SELECTION COMPUTED : Path, Stability selection, Lambda fixed
STABILITY SELECTION PARAMETERS: method = lam; lamin = 0.01; lam = theoritical; B = 50; q = 10; percent_nS = 0.5; threshold = 0.7; numerical_method = ODE
LAMBDA FIXED PARAMETERS: lam = theoritical; theoritical_lam = 19.1991; numerical_method = ODE
PATH PARAMETERS: Npath = 500 n_active = False lamin = 0.05 n_lam = 500; numerical_method = ODE
SIGMA FOR LAMFIXED : 0.7473015322224758
SPEEDNESS :
Running time for Path computation : 0.08s
Running time for Cross Validation : 'not computed'
Running time for Stability Selection : 1.374s
Running time for Fixed LAM : 0.024s
Optimization schemes
The available problem formulations [R1-C2] require different algorithmic strategies for efficiently solving the underlying optimization problem. We have implemented four algorithms (with provable convergence guarantees) that vary in generality and are not necessarily applicable to all problems. For each problem type, c-lasso has a default algorithm setting that proved to be the fastest in our numerical experiments.
Path algorithms (Path-Alg)
This is the default algorithm for non-concomitant problems [R1,R3,C1,C2]. The algorithm uses the fact that the solution path along λ is piecewise- affine (as shown, e.g., in [1]). When Least-Squares is used as objective function, we derive a novel efficient procedure that allows us to also derive the solution for the concomitant problem [R2] along the path with little extra computational overhead.
Projected primal-dual splitting method (P-PDS):
This algorithm is derived from [2] and belongs to the class of proximal splitting algorithms. It extends the classical Forward-Backward (FB) (aka proximal gradient descent) algorithm to handle an additional linear equality constraint via projection. In the absence of a linear constraint, the method reduces to FB. This method can solve problem [R1]. For the Huber problem [R3], P-PDS can solve the mean-shift formulation of the problem (see [6]).
Projection-free primal-dual splitting method (PF-PDS):
This algorithm is a special case of an algorithm proposed in [3] (Eq.4.5) and also belongs to the class of proximal splitting algorithms. The algorithm does not require projection operators which may be beneficial when C has a more complex structure. In the absence of a linear constraint, the method reduces to the Forward-Backward-Forward scheme. This method can solve problem [R1]. For the Huber problem [R3], PF-PDS can solve the mean-shift formulation of the problem (see [6]).
Douglas-Rachford-type splitting method (DR)
This algorithm is the most general algorithm and can solve all regression problems [R1-R4]. It is based on Doulgas Rachford splitting in a higher-dimensional product space. It makes use of the proximity operators of the perspective of the LS objective (see [4,5]) The Huber problem with concomitant scale [R4] is reformulated as scaled Lasso problem with the mean shift (see [6]) and thus solved in (n + d) dimensions.
Structure of the code
References
-
[1] B. R. Gaines, J. Kim, and H. Zhou, Algorithms for Fitting the Constrained Lasso, J. Comput. Graph. Stat., vol. 27, no. 4, pp. 861–871, 2018.
-
[2] L. Briceno-Arias and S.L. Rivera, A Projected Primal–Dual Method for Solving Constrained Monotone Inclusions, J. Optim. Theory Appl., vol. 180, Issue 3, March 2019.
-
[3] P. L. Combettes and J.C. Pesquet, Primal-Dual Splitting Algorithm for Solving Inclusions with Mixtures of Composite, Lipschitzian, and Parallel-Sum Type Monotone Operators, Set-Valued and Variational Analysis, vol. 20, pp. 307-330, 2012.
-
[4] P. L. Combettes and C. L. Müller, Perspective M-estimation via proximal decomposition, Electronic Journal of Statistics, 2020, Journal version
-
[5] P. L. Combettes and C. L. Müller, Regression models for compositional data: General log-contrast formulations, proximal optimization, and microbiome data applications, Statistics in Bioscience, 2020.
-
[6] A. Mishra and C. L. Müller, Robust regression with compositional covariates, arXiv, 2019.
-
[7] S. Rosset and J. Zhu, Piecewise linear regularized solution paths, Ann. Stat., vol. 35, no. 3, pp. 1012–1030, 2007.
Project details
Release history Release notifications | RSS feed
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.