import numpy as np
from sklearn.gaussian_process.kernels import Matern, RBF
from sklearn.ensemble import RandomForestRegressor

import plotly
import plotly.express as px

from docs.mse_estimator import ErrorComparer
from docs.data_generation import gen_rbf_X, gen_matern_X, gen_cov_mat
from spe.forest import ParametricRandomForestRegressor
from spe.tree import Tree
from spe.estimators import new_y_est, cp_bagged

np.random.seed(1)

Random Forest Comparisons

Here we compare the MSE of the usual parametric random forest (RF) and the non-parametric RF which is technically the model spe.estimators.cp_bagged.

## number of realizations to run
niter = 100

## data generation parameters
gsize=20
n=20**2
p=30
s=30
delta = 0.75
snr = 0.4
tr_frac = .5

noise_kernel = 'matern'
noise_length_scale = 1.
noise_nu = .5

X_kernel = 'matern'
X_length_scale = 5.
X_nu = 2.5

## ErrorComparer parameters
max_depth = 3
models = [
    [
        RandomForestRegressor(max_depth=max_depth, max_features='sqrt'),
        ParametricRandomForestRegressor(max_depth=max_depth, max_features='sqrt'),
        Tree(max_depth=max_depth, max_features='sqrt'),
    ]
]
ests = [
    new_y_est,
    new_y_est,
    cp_bagged,
]
est_kwargs = [
    {'alpha': None,
    'full_refit': False},
    {'alpha': None},
    {
        'use_trace_corr': False,
    },
]

## plot parameters
model_names = ["Blurred Forest"]
est_names = ["NPRF", "PRF", "GenCpPRF"]

err_cmp = ErrorComparer()

nx = ny = int(np.sqrt(n))
xs = np.linspace(0, gsize, nx)
ys = np.linspace(0, gsize, ny)
c_x, c_y = np.meshgrid(xs, ys)
c_x = c_x.flatten()
c_y = c_y.flatten()
coord = np.stack([c_x, c_y]).T

if noise_kernel == 'rbf':
    Sigma_t = gen_cov_mat(c_x, c_y, RBF(length_scale=noise_length_scale))
elif noise_kernel == 'matern':
    Sigma_t = gen_cov_mat(c_x, c_y, Matern(length_scale=noise_length_scale, nu=noise_nu))
else:
    Sigma_t = np.eye(n)

Cov_y_ystar = delta*Sigma_t
Sigma_t = delta*Sigma_t + (1-delta)*np.eye(n)

if noise_kernel == 'rbf' or noise_kernel == 'matern':
    Chol_y = np.linalg.cholesky(Sigma_t)
else:
    Chol_y = np.eye(n)

if X_kernel == 'rbf':
    X = gen_rbf_X(c_x, c_y, p)
elif X_kernel == 'matern':
    X = gen_matern_X(c_x, c_y, p, length_scale=X_length_scale, nu=X_nu)
else:
    X = np.random.randn(n,p)

beta = np.zeros(p)
idx = np.random.choice(p,size=s,replace=False)
beta[idx] = np.random.uniform(-1,1,size=s)

tr_idx = np.ones(n, dtype=bool)

model_errs = []

for model in models:
    errs = err_cmp.compare(
        model,
        ests,
        est_kwargs,
        niter=niter,
        n=n,
        p=p,
        s=s,
        snr=snr,
        X=X,
        beta=beta,
        coord=coord,
        Chol_y=Chol_y,
        Chol_ystar=None,
        Cov_y_ystar=None,
        tr_idx=tr_idx,
        fair=False,
    )
    model_errs.append(errs)

100%|██████████| 100/100 [01:03<00:00,  1.59it/s]

from importlib import reload
import docs.plotting_utils
reload(docs.plotting_utils)
from docs.plotting_utils import gen_model_barplots

plotly.offline.init_notebook_mode()
fig = gen_model_barplots(
    model_errs,
    model_names,
    est_names,
    title="RFs: Spatial Train/Test Split, SSN",
    has_test_risk=False,
    err_bars=True,
    # color_discrete_sequence=px.colors.qualitative.Bold[-1],
    color_discrete_sequence=px.colors.qualitative.Pastel[0],
    fig_name="p_vs_np_rf",
)
fig.show()