Skip to content

Instantly share code, notes, and snippets.

@stharrold

stharrold/20160124T213000Z_test_empirical_distribution_from_parent.ipynb

Created February 12, 2016 20:40
Show Gist options

  • Select an option

    Learn more about clone URLs
  • Save stharrold/e159a1eee2e69b12d66f to your computer and use it in GitHub Desktop.

Select an option

Learn more about clone URLs
Save stharrold/e159a1eee2e69b12d66f to your computer and use it in GitHub Desktop.
Download ZIP
20160124T213000Z_test_empirical_distribution_from_parent.ipynb
Display the source blob
Display the rendered blob
Raw
20160124T213000Z_test_empirical_distribution_from_parent.ipynb
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Test that an empirical distribution is from a parent distribution\n",
"\n",
"**Note:**\n",
"* This test is stricter than the Kolmogorov-Smirnov test (e.g. a p-value from K-S test of 0.09 would be 0.04 by this test).\n",
"* This test is stricter than the Anderson-Darling test (e.g. a p-value from A-D test of 0.09 would be 0.05 by this test).\n",
"\n",
"## Initialization\n",
"\n",
"### Imports"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"/home/samuel_harrold\n"
]
}
],
"source": [
"cd ~"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {
"collapsed": false
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/samuel_harrold/anaconda3/lib/python3.5/site-packages/sklearn/utils/fixes.py:64: DeprecationWarning: inspect.getargspec() is deprecated, use inspect.signature() instead\n",
" if 'order' in inspect.getargspec(np.copy)[0]:\n"
]
}
],
"source": [
"# Import standard packages.\n",
"import collections\n",
"import copy\n",
"import itertools\n",
"import operator\n",
"import os\n",
"import pdb\n",
"# Import installed packages.\n",
"import matplotlib.pyplot as plt\n",
"import nltk\n",
"import numpy as np\n",
"import pandas as pd\n",
"import scipy\n",
"import seaborn as sns\n",
"# Import local packages.\n",
"%reload_ext autoreload\n",
"%autoreload 2\n",
"%matplotlib inline"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Globals"
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"np.random.seed(seed=0)\n",
"sns.set()"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"def calc_prob_d1_from_d2(d1:np.ndarray, d2:np.ndarray, plot:bool=False) -> float:\n",
" r\"\"\"Calculate the probability that one sample distribution was drawn from another.\n",
" \n",
" Args:\n",
" d1 (numpy.ndarray): Sample to test as child distribution.\n",
" d2 (numpy.ndarray): Sample to test as parent distribution.\n",
" plot (bool, optional, False): Whether or not to create plots.\n",
" \n",
" Returns:\n",
" pval (float): P-value; probability that `d1` was drawn from `d2`.\n",
" See \"Notes\".\n",
" \n",
" Notes:\n",
" * `pval` is calculated as the fraction of subsamples from `d2`\n",
" that have more points than `d1` exceeding a confidence interval\n",
" about `d2` in a probability plot.\n",
" * The probability mass function of the number of values out of the CI\n",
" is an extreme value distribution.\n",
" \n",
" \"\"\"\n",
" # Compute percentiles...\n",
" percentiles = np.linspace(start=0.0, stop=100.0, num=101, endpoint=True)\n",
" # ...for d1\n",
" pctls_d1 = np.empty((len(percentiles), 2))\n",
" pctls_d1[:, 0] = percentiles\n",
" pctls_d1[:, 1] = np.percentile(d1, q=percentiles)\n",
" # ...for d2.\n",
" pctls_d2 = np.empty((len(percentiles), 2))\n",
" pctls_d2[:, 0] = percentiles\n",
" pctls_d2[:, 1] = np.percentile(d2, q=percentiles)\n",
" # ...for subsamples of d2.\n",
" subsamples = list()\n",
" nsubsamples = 1000\n",
" for _ in range(nsubsamples):\n",
" # TODO: For testing against a continuous distribution, replace `subsample`\n",
" # here with a continuous distribution.\n",
" subsample = np.random.choice(d2, size=len(d1), replace=True)\n",
" subsamples.append(subsample)\n",
" subsamples = np.asarray(subsamples)\n",
" pctls_subsamples = np.empty((len(percentiles), nsubsamples+1))\n",
" pctls_subsamples[:, 0] = percentiles\n",
" pctls_subsamples[:, 1:] = np.percentile(subsamples, q=percentiles, axis=1)\n",
" # ...for CI.\n",
" ci = [15.87, 84.13]\n",
" pctls_subsamples_ci = np.empty((len(percentiles), len(ci)+1))\n",
" pctls_subsamples_ci[:, 0] = percentiles\n",
" pctls_subsamples_ci[:, 1:] = np.percentile(pctls_subsamples[:, 1:], q=ci, axis=1).transpose()\n",
" # Find values from samples that exceed CI.\n",
" tfmask_d2_outci = np.logical_or(\n",
" pctls_d2[:, 1] < pctls_subsamples_ci[:, 1],\n",
" pctls_subsamples_ci[:, 2] < pctls_d2[:, 1])\n",
" tfmask_d1_outci = np.logical_or(\n",
" pctls_d1[:, 1] < pctls_subsamples_ci[:, 1],\n",
" pctls_subsamples_ci[:, 2] < pctls_d1[:, 1])\n",
" nvals_subsamples_outci = list()\n",
" for subsample in pctls_subsamples.transpose()[1:]:\n",
" nvals_subsamples_outci.append(np.sum(np.logical_or(\n",
" subsample < pctls_subsamples_ci[:, 1],\n",
" pctls_subsamples_ci[:, 2] < subsample)))\n",
" nvals_subsamples_outci = np.asarray(nvals_subsamples_outci)\n",
" ref_frac_d2_subsamples_nvals = (100 - (ci[1]-ci[0]))/100\n",
" test_frac_d2_subsamples_nvals = np.sum(nvals_subsamples_outci)/np.product(np.shape(pctls_subsamples[:, 1:]))\n",
" if not np.isclose(ref_frac_d2_subsamples_nvals, test_frac_d2_subsamples_nvals, atol=2*np.mean(np.diff(percentiles))):\n",
" raise AssertionError(\n",
" (\"Program error. The number of values outside of CI must agree with the defined CI.\\n\" +\n",
" \"ref_frac_d2_subsamples_nvals = {ref}\\n\" +\n",
" \"test_frac_d2_subsamples_nvals = {test}\").format(\n",
" ref=ref_frac_d2_subsamples_nvals, test=test_frac_d2_subsamples_nvals))\n",
" # Compute p-value of d1 being drawn from d2 as fraction of samples\n",
" # with more values that exceed CI.\n",
" frac_d2_nvals = np.sum(\n",
" np.sum(tfmask_d2_outci) <= nvals_subsamples_outci)/len(\n",
" nvals_subsamples_outci)\n",
" frac_d1_nvals = np.sum(\n",
" np.sum(tfmask_d1_outci) <= nvals_subsamples_outci)/len(\n",
" nvals_subsamples_outci)\n",
" pval = frac_d1_nvals\n",
" if plot:\n",
" # Similar to probability plot from scipy.stats.probplot\n",
" plt.plot(\n",
" pctls_d1[:, 0], pctls_d1[:, 1],\n",
" marker='.', label='d1 (dark=out of CI)',\n",
" color=sns.color_palette()[0])\n",
" plt.plot(\n",
" pctls_d1[tfmask_d1_outci, 0],\n",
" pctls_d1[tfmask_d1_outci, 1],\n",
" marker='o', linestyle='', label=None,\n",
" color=sns.color_palette(palette='dark')[0])\n",
" plt.plot(\n",
" pctls_d2[:, 0], pctls_d2[:, 1],\n",
" marker='.', label='d2 (dark=out of CI)',\n",
" color=sns.color_palette()[1])\n",
" plt.plot(\n",
" pctls_d2[tfmask_d2_outci, 0],\n",
" pctls_d2[tfmask_d2_outci, 1],\n",
" marker='o', linestyle='', label=None,\n",
" color=sns.color_palette(palette='dark')[1])\n",
" plt.fill_between(\n",
" pctls_subsamples_ci[:, 0],\n",
" y1=pctls_subsamples_ci[:, 1],\n",
" y2=pctls_subsamples_ci[:, 2],\n",
" alpha=0.5, color=sns.color_palette()[1],\n",
" label='16-84th pctl CI from d2 subsamples')\n",
" plt.title(\"Values vs percentiles\")\n",
" plt.xlabel(\"Percentile\")\n",
" plt.ylabel(\"value\")\n",
" plt.legend(loc='upper left')\n",
" plt.show()\n",
" plt.axvline(\n",
" x=np.sum(tfmask_d1_outci),\n",
" color=sns.color_palette(palette='dark')[0],\n",
" label='nvals d1 out of CI')\n",
" plt.axvline(\n",
" x=np.sum(tfmask_d2_outci),\n",
" color=sns.color_palette(palette='dark')[1],\n",
" label='nvals d2 out of CI')\n",
" weights = np.ones_like(nvals_subsamples_outci)/len(nvals_subsamples_outci)\n",
" hist_kws = {'weights':weights, 'histtype':'stepfilled'}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" nvals_subsamples_outci, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws,\n",
" label='nvals subsample out of CI', color=sns.color_palette()[1])\n",
" plt.title(\"Probability mass function of number of values out of CI\")\n",
" plt.xlabel(\"Number of values out of CI\")\n",
" plt.ylabel(\"P(nvals)\")\n",
" plt.legend(loc='upper left')\n",
" plt.show()\n",
" print(\"Probability that `d2` was drawn from `d2`: {prob:.1%}\".format(\n",
" prob=frac_d2_nvals))\n",
" print(\"Probability that `d1` was drawn from `d2`: {prob:.1%}\".format(\n",
" prob=frac_d1_nvals))\n",
" return pval"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Compare empirical distributions"
]
},
{
"cell_type": "code",
"execution_count": 122,
"metadata": {
"collapsed": false
},
"outputs": [],
"source": [
"# Define parent and trial distributions\n",
"# Note: To standardize `dist`,\n",
"# dist = (dist-np.mean(dist))/np.std(dist)\n",
"# Note: KS, AD tests perform especially poorly for distributions:\n",
"# geometric,\n",
"(size_parent, size_trial) = (100, 30)\n",
"#parent = np.random.beta(a=1, b=1, size=size_parent)\n",
"#parent = np.random.binomial(n=100, p=0.5, size=size_parent)\n",
"#parent = np.random.chisquare(df=1, size=size_parent)\n",
"#parent = np.random.dirichlet(alpha=(1,1), size=size_parent)[:, 0]\n",
"#parent = np.random.exponential(scale=2.0, size=size_parent)\n",
"#parent = np.random.f(dfnum=1, dfden=1, size=size_parent)\n",
"#parent = np.random.gamma(shape=1.0, scale=2.0, size=size_parent)\n",
"#parent = np.random.geometric(p=0.5, size=size_parent)\n",
"#parent = np.random.gumbel(loc=0.0, scale=1.0, size=size_parent)\n",
"#parent = np.random.hypergeometric(ngood=15, nbad=15, nsample=15, size=size_parent)\n",
"#parent = np.random.laplace(loc=0.0, scale=1.0, size=size_parent)\n",
"#parent = np.random.logistic(loc=0.0, scale=1.0, size=size_parent)\n",
"#parent = np.random.lognormal(mean=0.0, sigma=1.0, size=size_parent)\n",
"#parent = np.random.logseries(p=0.5, size=size_parent)\n",
"#parent = np.random.multinomial(n=20, pvals=[1/6]*6, size=size_parent)[:, 0]\n",
"#parent = np.random.multivariate_normal(mean=[0.0], cov=[[0.1]], size=size_parent)[:, 0]\n",
"#parent = np.random.negative_binomial(n=1, p=0.1, size=size_parent)\n",
"#parent = np.random.noncentral_chisquare(df=1, nonc=1.0, size=size_parent)\n",
"#parent = np.random.noncentral_f(dfnum=2, dfden=2, nonc=1, size=size_parent)\n",
"parent = np.random.normal(loc=1.0, scale=2.0, size=size_parent)\n",
"#parent = np.random.pareto(a=1.0, size=size_parent)\n",
"#parent = np.random.poisson(lam=1.0, size=size_parent)\n",
"#parent = np.random.power(a=3, size=size_parent)\n",
"#parent = np.random.rayleigh(scale=1.0, size=size_parent)\n",
"#parent = np.random.standard_cauchy(size=size_parent)\n",
"#parent = np.random.standard_exponential(size=size_parent)\n",
"#parent = np.random.standard_gamma(shape=1.0, size=size_parent)\n",
"#parent = np.random.standard_normal(size=size_parent)\n",
"#parent = np.random.standard_t(df=10, size=size_parent)\n",
"#parent = np.random.triangular(left=-1.0, mode=0.0, right=1.0, size=size_parent)\n",
"#parent = np.random.uniform(low=0.0, high=1.0, size=size_parent)\n",
"#parent = np.random.vonmises(mu=0.0, kappa=np.pi, size=size_parent)\n",
"#parent = np.random.wald(mean=1.0, scale=1.0, size=size_parent)\n",
"#parent = np.random.weibull(a=5.0, size=size_parent)\n",
"#parent = np.random.zipf(a=2.0, size=size_parent)\n",
"trial = np.random.choice(parent, size=size_trial, replace=True)"
]
},
{
"cell_type": "code",
"execution_count": 123,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/samuel_harrold/anaconda3/lib/python3.5/site-packages/scipy/stats/morestats.py:1438: UserWarning: approximate p-value will be computed by extrapolation\n",
" warnings.warn(\"approximate p-value will be computed by extrapolation\")\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfoAAAFwCAYAAABQPrn3AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79f7946b38>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Fraction of tests with p-values <= 0.05: (name, frac)\n",
"(emp, 0.05)\n",
"(ks, 0.01)\n",
"(ad, 0.03)\n",
"\n",
"p-values for `trial` sample closest to p-value=0.05\n"
]
},
{
"data": {
"text/plain": [
"emp 0.052000\n",
"ks 0.086481\n",
"ad 0.171799\n",
"Name: 63, dtype: float64"
]
},
"execution_count": 123,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# TEST: Compare to other tests\n",
"# For chi2 test; f_obs, f_exp must be about the same sample size.\n",
"# Other tests can use different sized `parent` and `trial`.\n",
"pvals = list()\n",
"idx_pvals05_trial = None\n",
"for idx in range(100):\n",
" trial = np.random.choice(parent, size=size_trial, replace=True)\n",
" pvals_trial = list()\n",
" # Empirical test\n",
" pval_emp = calc_prob_d1_from_d2(d1=trial, d2=parent)\n",
" pvals_trial.append(pval_emp)\n",
" # # TODO: Probably not computing Chi2 test correctly for different sized samples.\n",
" # # Chi2 test.\n",
" # expected = np.random.choice(parent, size=size_trial, replace=True)\n",
" # bins = sns.distributions._freedman_diaconis_bins(expected)\n",
" # (f_exp, bin_edges) = np.histogram(expected, bins=bins)\n",
" # (f_obs, _) = np.histogram(trial, bins=bin_edges)\n",
" # pval_chi2 = scipy.stats.chisquare(f_obs=f_obs, f_exp=f_exp, ddof=0)[-1]\n",
" # pvals_trial.append(pval_chi2)\n",
" # Kolmogorov-Smirnov test\n",
" pval_ks = scipy.stats.ks_2samp(data1=parent, data2=trial)[-1]\n",
" pvals_trial.append(pval_ks)\n",
" # Anderson-Darling test\n",
" pval_ad = scipy.stats.anderson_ksamp(samples=[parent, trial])[-1]\n",
" pvals_trial.append(pval_ad)\n",
" # # TODO: Probably shouldn't combine p-values since evaluating independent methods.\n",
" # # Combined p-values, excluding the p-value from the empirical test.\n",
" # pval_comb = scipy.stats.combine_pvalues(\n",
" # pvalues=pvals_trial[1:], method='stouffer', weights=None)[-1]\n",
" # pvals_trial.append(pval_comb)\n",
" # Save p-values and find border-line significant distribution.\n",
" pvals.append(pvals_trial)\n",
" if (idx_pvals05_trial is None\n",
" or abs(0.05-pval_emp) < abs(0.05-idx_pvals05_trial[1])):\n",
" idx_pvals05_trial = (idx, *pvals_trial, trial)\n",
"df_pvals = pd.DataFrame.from_records(pvals, columns=['emp', 'ks', 'ad'])\n",
"\n",
"# Plot distribution tests.\n",
"xvals = df_pvals['emp'].values\n",
"col_plotkwargs = {\n",
" 'emp': {'marker':'.'},\n",
" 'comb': {'marker':'o'}}\n",
"for col in df_pvals.columns.values:\n",
" plotkwargs = col_plotkwargs.get(\n",
" col, {'marker':'.'})\n",
" plt.plot(\n",
" xvals, df_pvals[col].values, label=col,\n",
" linestyle='', **plotkwargs)\n",
"plt.title(\"Distribution test p-values vs empirical test p-values\")\n",
"plt.xlabel(\"Empirical test\")\n",
"plt.ylabel(\"Other test\")\n",
"plt.legend(loc='upper left')\n",
"plt.show()\n",
"print(\"Fraction of tests with p-values <= 0.05: (name, frac)\")\n",
"for col in df_pvals.columns.values:\n",
" frac = np.sum(df_pvals[col] <= 0.05)/len(df_pvals)\n",
" print(\"({name}, {frac:.2f})\".format(name=col, frac=frac))\n",
"print()\n",
"\n",
"# Set `trial` to sample closest to p-value=0.05.\n",
"print(\"p-values for `trial` sample closest to p-value=0.05\")\n",
"trial = idx_pvals05_trial[-1]\n",
"assert np.all(np.isclose(idx_pvals05_trial[1:-1], df_pvals.iloc[idx_pvals05_trial[0]]))\n",
"df_pvals.iloc[idx_pvals05_trial[0]]"
]
},
{
"cell_type": "code",
"execution_count": 124,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"################################################################################\n",
"Plot parent\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAFwCAYAAACy4qKOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79f7929da0>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFwCAYAAAChNeJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nGOL/Ayk3kSKOPqJfAAAAAElFTkSuQmCC\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79f9d21390>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"################################################################################\n",
"Plot trial\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAf0AAAFwCAYAAACy4qKOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79fa3cd668>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFwCAYAAAChNeJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nRK5CYII=\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79f9ce56a0>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"################################################################################\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfsAAAFwCAYAAAC//NLJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79fa8112e8>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAgIAAAFwCAYAAADKRxObAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nTkSuQmCC\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f79fa23fa90>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Probability that `d2` was drawn from `d2`: 100.0%\n",
"Probability that `d1` was drawn from `d2`: 6.3%\n"
]
},
{
"data": {
"text/plain": [
"0.063"
]
},
"execution_count": 124,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Plot distributions and comparison.\n",
"for (name, dist) in [('parent', parent), ('trial', trial)]:\n",
" print('#'*80)\n",
" print(\"Plot {name}\".format(name=name))\n",
" weights = np.ones_like(dist)/len(dist)\n",
" hist_kws = {'weights':weights, 'histtype':'stepfilled'}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws)\n",
" plt.title('Probability density function')\n",
" plt.ylabel('P(v_lwr <= v < v_upr)')\n",
" plt.xlabel('Value')\n",
" plt.show()\n",
" hist_kws = {\n",
" 'weights':weights, 'histtype':'step',\n",
" 'cumulative':True, 'linewidth':1, 'alpha':0.5}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws)\n",
" plt.title('Cumulative distribution function')\n",
" plt.xlabel('Value')\n",
" plt.ylabel('P(v < value)')\n",
" plt.show()\n",
"print('#'*80)\n",
"calc_prob_d1_from_d2(d1=trial, d2=parent, plot=True)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Example from scipy.stats.ks_2samp\n",
"\n",
"http://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.stats.ks_2samp.html"
]
},
{
"cell_type": "code",
"execution_count": 130,
"metadata": {
"collapsed": true
},
"outputs": [],
"source": [
"from scipy import stats\n",
"np.random.seed(12345678) #fix random seed to get the same result\n",
"n1 = 200 # size of first sample\n",
"n2 = 300 # size of second sample"
]
},
{
"cell_type": "code",
"execution_count": 131,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"Ks_2sampResult(statistic=0.20833333333333337, pvalue=4.6674975515806989e-05)"
]
},
"execution_count": 131,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# For a different distribution, we can reject the null hypothesis since the pvalue is below 1%:\n",
"rvs1 = stats.norm.rvs(size=n1, loc=0., scale=1)\n",
"rvs2 = stats.norm.rvs(size=n2, loc=0.5, scale=1.5)\n",
"stats.ks_2samp(rvs1, rvs2)"
]
},
{
"cell_type": "code",
"execution_count": 132,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"Ks_2sampResult(statistic=0.10333333333333333, pvalue=0.14498781825751686)"
]
},
"execution_count": 132,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"#For a slightly different distribution, we cannot reject the null hypothesis at a 10%\n",
"# or lower alpha since the p-value at 0.144 is higher than 10%\n",
"rvs3 = stats.norm.rvs(size=n2, loc=0.01, scale=1.0)\n",
"stats.ks_2samp(rvs1, rvs3)"
]
},
{
"cell_type": "code",
"execution_count": 133,
"metadata": {
"collapsed": false
},
"outputs": [
{
"data": {
"text/plain": [
"Ks_2sampResult(statistic=0.07999999999999996, pvalue=0.41126949729859719)"
]
},
"execution_count": 133,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# For an identical distribution, we cannot reject the null hypothesis since the p-value is high, 41%:\n",
"rvs4 = stats.norm.rvs(size=n2, loc=0.0, scale=1.0)\n",
"stats.ks_2samp(rvs1, rvs4)"
]
},
{
"cell_type": "code",
"execution_count": 137,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Is the probability low that the rvs1 was drawn from rvs2 and vice-versa?\n"
]
},
{
"data": {
"text/plain": [
"True"
]
},
"execution_count": 137,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print(\"Is the probability low that the rvs1 was drawn from rvs2 and vice-versa?\")\n",
"(calc_prob_d1_from_d2(d1=rvs1, d2=rvs2) < 0.05) and (calc_prob_d1_from_d2(d1=rvs2, d2=rvs1) < 0.05)"
]
},
{
"cell_type": "code",
"execution_count": 138,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Is the probability low that the rvs1 was drawn from rvs3 and vice-versa?\n"
]
},
{
"data": {
"text/plain": [
"False"
]
},
"execution_count": 138,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print(\"Is the probability low that the rvs1 was drawn from rvs3 and vice-versa?\")\n",
"(calc_prob_d1_from_d2(d1=rvs1, d2=rvs3) < 0.05) and (calc_prob_d1_from_d2(d1=rvs3, d2=rvs1) < 0.05)"
]
},
{
"cell_type": "code",
"execution_count": 139,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Is the probability low that the rvs1 was drawn from rvs4 and vice-versa?\n"
]
},
{
"data": {
"text/plain": [
"True"
]
},
"execution_count": 139,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print(\"Is the probability low that the rvs1 was drawn from rvs4 and vice-versa?\")\n",
"(calc_prob_d1_from_d2(d1=rvs1, d2=rvs4) < 0.05) and (calc_prob_d1_from_d2(d1=rvs4, d2=rvs1) < 0.05)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Example for scipy.stats.anderson_ksamp\n",
"\n",
"http://docs.scipy.org/doc/scipy-0.14.0/reference/generated/scipy.stats.anderson_ksamp.html"
]
},
{
"cell_type": "code",
"execution_count": 144,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stderr",
"output_type": "stream",
"text": [
"/home/samuel_harrold/anaconda3/lib/python3.5/site-packages/scipy/stats/morestats.py:1438: UserWarning: approximate p-value will be computed by extrapolation\n",
" warnings.warn(\"approximate p-value will be computed by extrapolation\")\n"
]
},
{
"data": {
"text/plain": [
"Anderson_ksampResult(statistic=10.997724734808768, critical_values=array([ 0.325, 1.226, 1.961, 2.718, 3.752]), significance_level=7.979896693422518e-05)"
]
},
"execution_count": 144,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"stats.anderson_ksamp(samples=[rvs1, rvs2])"
]
},
{
"cell_type": "code",
"execution_count": 145,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"data": {
"text/plain": [
"Anderson_ksampResult(statistic=0.35740095935945254, critical_values=array([ 0.325, 1.226, 1.961, 2.718, 3.752]), significance_level=0.2408137175052366)"
]
},
"execution_count": 145,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"stats.anderson_ksamp(samples=[rvs1, rvs3])"
]
},
{
"cell_type": "code",
"execution_count": 146,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"data": {
"text/plain": [
"Anderson_ksampResult(statistic=1.3133061476271048, critical_values=array([ 0.325, 1.226, 1.961, 2.718, 3.752]), significance_level=0.09276304371769672)"
]
},
"execution_count": 146,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"stats.anderson_ksamp(samples=[rvs1, rvs4])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## ARCHIVED"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {
"collapsed": false,
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"################################################################################\n",
"Plot parent\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfYAAAFwCAYAAABKAll5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f4168193d30>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFwCAYAAAChNeJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nCEodAABD/H88RgkXxKAHCwAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f414849de10>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"################################################################################\n",
"Plot experiment\n"
]
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfYAAAFwCAYAAABKAll5AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f41482cbdd8>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAfUAAAFwCAYAAAChNeJ6AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nDgCAISh1AAAMQakDAGCI/weJ5g8BDZzuTQAAAABJRU5ErkJggg==\n",
"text/plain": [
"<matplotlib.figure.Figure at 0x7f414821cc50>"
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"# Plot parent distribution\n",
"for (name, dist) in [('parent', parent), ('experiment', experiment)]:\n",
" print('#'*80) \n",
" print(\"Plot {name}\".format(name=name))\n",
" weights = np.ones_like(dist)/len(dist)\n",
" # Note: seaborn can't broadcast some discrete distributions\n",
" # https://github.com/stharrold/sandbox/issues/6\n",
" try:\n",
" hist_kws = {'weights':weights, 'histtype':'stepfilled'}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws)\n",
" except ValueError:\n",
" bin_width = (np.max(dist)-np.min(dist))/(len(np.unique(dist))-1)\n",
" bin_edges=np.linspace(\n",
" start=np.min(dist), stop=np.max(dist)+bin_width,\n",
" num=len(np.unique(dist))+1, endpoint=True)\n",
" hist_kws = {\n",
" 'weights':weights, 'histtype':'step', 'align':'left',\n",
" 'cumulative':False, 'linewidth':1, 'alpha':0.5}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, bins=bin_edges, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws, color=sns.color_palette()[0])\n",
" plt.title('Probability density function')\n",
" plt.ylabel('P(v_lwr <= v < v_upr)')\n",
" plt.xlabel('Value')\n",
" plt.show()\n",
" # Note: seaborn can't broadcast some discrete distributions\n",
" # https://github.com/stharrold/sandbox/issues/6\n",
" try:\n",
" hist_kws = {\n",
" 'weights':weights, 'histtype':'step',\n",
" 'cumulative':True, 'linewidth':1, 'alpha':0.5}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws)\n",
" except ValueError:\n",
" bin_width = (np.max(dist)-np.min(dist))/(len(np.unique(dist))-1)\n",
" bin_edges=np.linspace(\n",
" start=np.min(dist), stop=np.max(dist)+bin_width,\n",
" num=len(np.unique(dist))+1, endpoint=True)\n",
" hist_kws = {\n",
" 'weights':weights, 'histtype':'step', 'align':'left',\n",
" 'cumulative':True, 'linewidth':1, 'alpha':0.5}\n",
" rug_kws = {'alpha':0.5}\n",
" sns.distplot(\n",
" dist, bins=bin_edges, kde=False, rug=True, norm_hist=False,\n",
" hist_kws=hist_kws, rug_kws=rug_kws, color=sns.color_palette()[0]) \n",
" plt.title('Cumulative distribution function')\n",
" plt.xlabel('Value')\n",
" plt.ylabel('P(v < value)')\n",
" plt.show()"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.5.1"
}
},
"nbformat": 4,
"nbformat_minor": 0
}
@stharrold
Copy link
Author

stharrold commented Feb 28, 2016

TODO:

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment