File size: 11,606 Bytes
798602c | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 | from __future__ import annotations
from typing import Optional, Tuple
import numpy as np
import pandas as pd
from core.estimation.inference.estimators import estimate_mean, estimate_sigma
from core.estimation.inference.ci import (
ci_mean_analytic,
ci_mean_bootstrap,
ci_median_analytic,
ci_median_bootstrap,
)
from core.estimation.inference.pi import (
pi_mean,
pi_median,
pi_iqr,
pi_bootstrap,
)
from core.estimation.graphical_analysis import (
plot_histogram_with_overlays,
plot_ecdf,
)
# ---------------------------------------------------------------------
# Utilities (aligned with inference_controller)
# ---------------------------------------------------------------------
def select_distribution(mean_estimator: str, sigma_estimator: str) -> str:
if (
mean_estimator == "Sample Mean"
and sigma_estimator == "Deviation (1 ddof)"
):
return "t"
return "norm"
def validate_deviation_estimator(*, sigma_estimator: str, n: int):
if sigma_estimator == "Range (bias corrected)" and n > 25:
raise ValueError(
"Range-based confidence intervals require n ≤ 25. "
"Use another estimator or bootstrap."
)
def _prepare_series(
df: pd.DataFrame,
column: str,
weights_col: Optional[str],
) -> tuple[np.ndarray, Optional[np.ndarray]]:
if df is None:
raise ValueError("No data loaded. Please load a dataset first.")
if column not in df.columns:
raise ValueError(f"Column '{column}' not found in the dataframe.")
series = df[column].dropna()
if series.empty:
raise ValueError(f"Column '{column}' has no non-missing values.")
weights = None
if weights_col is not None:
if weights_col not in df.columns:
raise ValueError(
f"Weights column '{weights_col}' not found in the dataframe."
)
weights_series = df[weights_col].reindex(series.index).dropna()
common_idx = series.index.intersection(weights_series.index)
series = series.loc[common_idx]
weights_series = weights_series.loc[common_idx]
weights = weights_series.to_numpy()
return series.to_numpy(), weights
def run_graphical_analysis(
*,
df: pd.DataFrame,
column: str,
graph_type: str,
# Histogram / PMF controls
add_kde: bool,
add_data: bool,
add_normal: bool,
add_ci: bool,
ci_choice: str,
add_pi: bool,
pi_choice: str,
# Estimators
mean_estimator: str,
median_estimator: str,
sigma_estimator: str,
trim_param,
winsor_limits,
weights_col: Optional[str],
# Normal μ source
normal_mu_source: str,
# Bootstrap options
bootstrap_mean: bool,
bootstrap_median: bool,
bootstrap_sigma: bool,
bootstrap_prediction: bool,
bootstrap_samples: int,
# CI/PI confidence level
ci_pi_conf_level: float,
# ECDF controls
ecdf_add_conf: bool,
ecdf_conf_level: float,
ecdf_add_normal: bool,
):
data, weights = _prepare_series(df, column, weights_col)
if not (0.0 < ci_pi_conf_level < 1.0):
raise ValueError("Confidence level for CI/PI must be in (0, 1).")
if graph_type in ("Histogram", "Empirical Probability Mass Function"):
return _run_hist_or_pmf(
data=data,
var_name=column,
graph_type=graph_type,
add_kde=add_kde,
add_data=add_data,
add_normal=add_normal,
add_ci=add_ci,
ci_choice=ci_choice,
add_pi=add_pi,
pi_choice=pi_choice,
mean_estimator=mean_estimator,
median_estimator=median_estimator,
sigma_estimator=sigma_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
normal_mu_source=normal_mu_source,
bootstrap_mean=bootstrap_mean,
bootstrap_median=bootstrap_median,
bootstrap_sigma=bootstrap_sigma,
bootstrap_prediction=bootstrap_prediction,
bootstrap_samples=bootstrap_samples,
ci_pi_conf_level=ci_pi_conf_level,
)
if graph_type == "Empirical Cumulative Distribution Function (ECDF)":
return _run_ecdf(
data=data,
var_name=column,
ecdf_add_conf=ecdf_add_conf,
ecdf_conf_level=ecdf_conf_level,
ecdf_add_normal=ecdf_add_normal,
mean_estimator=mean_estimator,
sigma_estimator=sigma_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
normal_mu_source=normal_mu_source,
)
raise ValueError(f"Unknown graph type: {graph_type}")
def _run_hist_or_pmf(
*,
data: np.ndarray,
var_name: str,
graph_type: str,
add_kde: bool,
add_data: bool,
add_normal: bool,
add_ci: bool,
ci_choice: str,
add_pi: bool,
pi_choice: str,
mean_estimator: str,
median_estimator: str,
sigma_estimator: str,
trim_param,
winsor_limits,
weights: Optional[np.ndarray],
normal_mu_source: str,
bootstrap_mean: bool,
bootstrap_median: bool,
bootstrap_sigma: bool,
bootstrap_prediction: bool,
bootstrap_samples: int,
ci_pi_conf_level: float,
):
alpha = 1.0 - ci_pi_conf_level
n = len(data)
validate_deviation_estimator(
sigma_estimator=sigma_estimator,
n=n,
)
ci_mean_interval = None
ci_median_interval = None
pi_interval = None
hat_mu = None
hat_sigma = None
need_intervals = add_ci or add_pi or add_normal
if need_intervals:
# --- Parameters for Normal overlay ---
if add_normal:
if normal_mu_source == "Mean-based CI":
hat_mu = estimate_mean(
data,
mean_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
)
else:
hat_mu = float(np.median(data))
hat_sigma = estimate_sigma(
data=data,
estimator=sigma_estimator,
)
# --- Confidence intervals ---
if add_ci:
dist = select_distribution(mean_estimator, sigma_estimator)
# CI for mean
if bootstrap_mean:
ci_mean_interval = ci_mean_bootstrap(
data=data,
estimator=mean_estimator,
alpha=alpha,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
B=bootstrap_samples,
)
else:
ci_mean_interval = ci_mean_analytic(
data=data,
estimator=mean_estimator,
alpha=alpha,
dist=dist,
sigma_estimator=sigma_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
)
# CI for median
if bootstrap_median:
ci_median_interval = ci_median_bootstrap(
data=data,
alpha=alpha,
B=bootstrap_samples,
)
else:
ci_median_interval = ci_median_analytic(
data=data,
alpha=alpha,
sigma_estimator=sigma_estimator,
)
# Respect user choice (Mean / Median / Both)
if ci_choice == "Mean":
ci_median_interval = None
elif ci_choice == "Median":
ci_mean_interval = None
# --- Prediction intervals ---
if add_pi:
dist = select_distribution(mean_estimator, sigma_estimator)
if pi_choice == "Mean":
pi_interval = pi_mean(
data=data,
alpha=alpha,
estimator=mean_estimator,
dist=dist,
sigma_estimator=sigma_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
)
elif pi_choice == "Median":
# New API: pi_median only needs data, alpha and sigma_estimator
pi_interval = pi_median(
data=data,
alpha=alpha,
sigma_estimator=sigma_estimator,
)
elif pi_choice == "IQR":
pi_interval = pi_iqr(
data=data,
alpha=alpha,
)
elif pi_choice == "Bootstrap":
if not bootstrap_prediction:
raise ValueError(
"To use the Bootstrap prediction interval, enable the "
"'Bootstrap Prediction' option in the estimator settings."
)
pi_interval = pi_bootstrap(
data=data,
alpha=alpha,
B=bootstrap_samples,
)
else:
raise ValueError(
f"Unknown prediction-interval choice: {pi_choice}"
)
fig = plot_histogram_with_overlays(
data=data,
graph_type=graph_type,
var_name=var_name,
add_kde=add_kde,
add_data=add_data,
add_normal=add_normal,
hat_mu=hat_mu,
hat_sigma=hat_sigma,
ci_mean_interval=ci_mean_interval,
ci_median_interval=ci_median_interval,
pi_interval=pi_interval,
)
return fig
def _run_ecdf(
*,
data: np.ndarray,
var_name: str,
ecdf_add_conf: bool,
ecdf_conf_level: float,
ecdf_add_normal: bool,
mean_estimator: str,
sigma_estimator: str,
trim_param,
winsor_limits,
weights: Optional[np.ndarray],
normal_mu_source: str,
):
if not (0.0 < ecdf_conf_level < 1.0):
raise ValueError("ECDF confidence level must be in (0, 1).")
alpha = 1.0 - ecdf_conf_level
n = len(data)
validate_deviation_estimator(
sigma_estimator=sigma_estimator,
n=n,
)
hat_mu = None
hat_sigma = None
if ecdf_add_normal:
if normal_mu_source == "Mean-based CI":
hat_mu = estimate_mean(
data,
mean_estimator,
trim_param=trim_param,
winsor_limits=winsor_limits,
weights=weights,
)
else:
hat_mu = float(np.median(data))
hat_sigma = estimate_sigma(
data=data,
estimator=sigma_estimator,
)
fig = plot_ecdf(
data=data,
var_name=var_name,
alpha=alpha,
add_conf_band=ecdf_add_conf,
add_normal=ecdf_add_normal,
hat_mu=hat_mu,
hat_sigma=hat_sigma,
)
return fig
|