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) # %% # Remark : one can see the parameters that should be selected : print(np.nonzero(sol)) # %% # Define the classo instance # ^^^^^^^^^^^^^^^^^^^^^^^^^^^ # # Next we can define a default c-lasso problem instance with the generated data: problem = classo_problem(X, y, C) # %% # Check parameters # ^^^^^^^^^^^^^^^^^^^^^^^^^^^ # # You can look at the generated problem instance by typing: print(problem) # %% # Solve optimization problems # ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ # # We only use stability selection as default model selection strategy. # The command also allows you to inspect the computed stability profile for all variables
path = "../../figures/examplePH/"
from classo import classo_problem
import numpy as np
from copy import deepcopy as dc
import scipy.io as sio
pH = sio.loadmat("pH_data/matlab/pHData.mat")
tax = sio.loadmat("pH_data/matlab/taxTablepHData.mat")["None"][0]
X, Y_uncent = pH["X"], pH["Y"].T[0]
y = Y_uncent - np.mean(Y_uncent) # Center Y
problem = classo_problem(X, y) # zero sum is default C
# Solve the entire path
problem.model_selection.PATH = True
problem.solve()
problem.solution.PATH.save = path + "R3-"
problem.solution.StabSel.save1 = path + "R3-StabSel"
problem.solution.StabSel.save3 = path + "R3-StabSel-beta"
problem1 = dc(problem)
# problem.formulation.huber = True
problem.solve()
problem.solution.PATH.save = path + "R4-"
problem.solution.StabSel.save1 = path + "R4-StabSel"
problem.solution.StabSel.save3 = path + "R4-StabSel-beta"
problem2 = dc(problem)
print(problem1, problem1.solution)
print(problem2, problem2.solution)
label_short = np.array([l.split("::")[-1] for l in label])
pseudo_count = 1
X = np.log(pseudo_count + x)
nleaves = np.sum(A, axis=0)
logGeom = X.dot(A) / nleaves
n, d = logGeom.shape
tr = np.random.permutation(n)[:int(0.8 * n)]
# %%
# Cross validation and Path Computation
# ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
problem = classo_problem(logGeom[tr], y[tr], label=label_short)
problem.formulation.w = 1 / nleaves
problem.formulation.intercept = True
problem.formulation.concomitant = False
problem.model_selection.StabSel = False
problem.model_selection.PATH = True
problem.model_selection.CV = True
problem.model_selection.CVparameters.seed = 6 # one could change logscale, Nsubset, oneSE
print(problem)
problem.solve()
print(problem.solution)
selection = problem.solution.CV.selected_param[1:] # exclude the intercept
# %%
# 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) # Joint microbiome and covariate data and offset
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 stability selection with theoretical lambda [Combettes & Müller, 2020b]
problem.model_selection.StabSelparameters.method = 'lam'
problem.model_selection.StabSelparameters.threshold_label = 0.5
# %%
# Use formulation R3
# ^^^^^^^^^^^^^^^^^^^^^^^^^^^
problem.formulation.concomitant = True
problem.solve()
print(problem)
print(problem.solution)
