better support for multivariate treatment in identification symbolic outputs
This commit is contained in:
Amit Sharma
Родитель
449fe0d34e
Коммит
532208e05d
|
|
@ -13,7 +13,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 24,
|
||||
"execution_count": 1,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
|
|
@ -23,7 +23,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 25,
|
||||
"execution_count": 2,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
|
|
@ -47,7 +47,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 26,
|
||||
"execution_count": 3,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -86,62 +86,62 @@
|
|||
" <tr>\n",
|
||||
" <th>0</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.977197</td>\n",
|
||||
" <td>-0.513249</td>\n",
|
||||
" <td>-1.085495</td>\n",
|
||||
" <td>-0.947936</td>\n",
|
||||
" <td>-0.836739</td>\n",
|
||||
" <td>-0.241959</td>\n",
|
||||
" <td>0.297495</td>\n",
|
||||
" <td>1.104464</td>\n",
|
||||
" <td>-1.243507</td>\n",
|
||||
" <td>2.078729</td>\n",
|
||||
" <td>-0.035286</td>\n",
|
||||
" <td>-1.763171</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>-0.443120</td>\n",
|
||||
" <td>11.701762</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>1</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.133648</td>\n",
|
||||
" <td>1.718109</td>\n",
|
||||
" <td>1.532397</td>\n",
|
||||
" <td>-0.441355</td>\n",
|
||||
" <td>0.338025</td>\n",
|
||||
" <td>1.321070</td>\n",
|
||||
" <td>0.940643</td>\n",
|
||||
" <td>1.552100</td>\n",
|
||||
" <td>0.374128</td>\n",
|
||||
" <td>-0.198048</td>\n",
|
||||
" <td>-0.202730</td>\n",
|
||||
" <td>0.024421</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>23.899661</td>\n",
|
||||
" <td>13.893867</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>2</th>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.261689</td>\n",
|
||||
" <td>2.167228</td>\n",
|
||||
" <td>-0.133905</td>\n",
|
||||
" <td>-0.641217</td>\n",
|
||||
" <td>1.669584</td>\n",
|
||||
" <td>0.555210</td>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.782048</td>\n",
|
||||
" <td>0.413752</td>\n",
|
||||
" <td>-0.404527</td>\n",
|
||||
" <td>1.963213</td>\n",
|
||||
" <td>1.579329</td>\n",
|
||||
" <td>0.539134</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>24.951954</td>\n",
|
||||
" <td>22.024896</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>3</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.699376</td>\n",
|
||||
" <td>-0.164822</td>\n",
|
||||
" <td>1.277675</td>\n",
|
||||
" <td>-1.814968</td>\n",
|
||||
" <td>-0.757798</td>\n",
|
||||
" <td>0.253259</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>8.803943</td>\n",
|
||||
" <td>0.542946</td>\n",
|
||||
" <td>-2.670037</td>\n",
|
||||
" <td>0.290079</td>\n",
|
||||
" <td>-0.104658</td>\n",
|
||||
" <td>-1.642732</td>\n",
|
||||
" <td>3.386802</td>\n",
|
||||
" <td>False</td>\n",
|
||||
" <td>6.253032</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>4</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.669321</td>\n",
|
||||
" <td>0.703854</td>\n",
|
||||
" <td>0.818505</td>\n",
|
||||
" <td>0.106033</td>\n",
|
||||
" <td>0.139291</td>\n",
|
||||
" <td>-0.996312</td>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.369742</td>\n",
|
||||
" <td>-0.503177</td>\n",
|
||||
" <td>0.160772</td>\n",
|
||||
" <td>2.177942</td>\n",
|
||||
" <td>-0.401280</td>\n",
|
||||
" <td>1.727678</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>15.721275</td>\n",
|
||||
" <td>25.160176</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>...</th>\n",
|
||||
|
|
@ -157,63 +157,63 @@
|
|||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>9995</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.952999</td>\n",
|
||||
" <td>0.983547</td>\n",
|
||||
" <td>1.205418</td>\n",
|
||||
" <td>-0.845736</td>\n",
|
||||
" <td>-1.665635</td>\n",
|
||||
" <td>-0.285645</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>9.965975</td>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.314375</td>\n",
|
||||
" <td>0.410457</td>\n",
|
||||
" <td>0.025014</td>\n",
|
||||
" <td>-1.307578</td>\n",
|
||||
" <td>0.495262</td>\n",
|
||||
" <td>0.247480</td>\n",
|
||||
" <td>False</td>\n",
|
||||
" <td>-2.947694</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>9996</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.893868</td>\n",
|
||||
" <td>0.147810</td>\n",
|
||||
" <td>0.320192</td>\n",
|
||||
" <td>-0.776280</td>\n",
|
||||
" <td>-0.458415</td>\n",
|
||||
" <td>-0.824306</td>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.134332</td>\n",
|
||||
" <td>-0.249336</td>\n",
|
||||
" <td>0.936425</td>\n",
|
||||
" <td>0.620813</td>\n",
|
||||
" <td>1.940425</td>\n",
|
||||
" <td>0.880728</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>8.452899</td>\n",
|
||||
" <td>18.983531</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>9997</th>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.818243</td>\n",
|
||||
" <td>0.422534</td>\n",
|
||||
" <td>-0.360710</td>\n",
|
||||
" <td>-0.761213</td>\n",
|
||||
" <td>-1.156205</td>\n",
|
||||
" <td>0.996336</td>\n",
|
||||
" <td>0.121186</td>\n",
|
||||
" <td>0.285559</td>\n",
|
||||
" <td>0.554069</td>\n",
|
||||
" <td>0.496691</td>\n",
|
||||
" <td>-0.324387</td>\n",
|
||||
" <td>-2.102712</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>5.366551</td>\n",
|
||||
" <td>4.754810</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>9998</th>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.485403</td>\n",
|
||||
" <td>0.665532</td>\n",
|
||||
" <td>1.046824</td>\n",
|
||||
" <td>-1.026499</td>\n",
|
||||
" <td>0.815819</td>\n",
|
||||
" <td>2.431939</td>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.400960</td>\n",
|
||||
" <td>-0.326070</td>\n",
|
||||
" <td>1.179469</td>\n",
|
||||
" <td>1.196163</td>\n",
|
||||
" <td>-0.216770</td>\n",
|
||||
" <td>-0.173123</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>20.121092</td>\n",
|
||||
" <td>15.567185</td>\n",
|
||||
" </tr>\n",
|
||||
" <tr>\n",
|
||||
" <th>9999</th>\n",
|
||||
" <td>0.0</td>\n",
|
||||
" <td>0.334394</td>\n",
|
||||
" <td>1.608105</td>\n",
|
||||
" <td>-1.329177</td>\n",
|
||||
" <td>-2.369921</td>\n",
|
||||
" <td>0.185035</td>\n",
|
||||
" <td>-0.092841</td>\n",
|
||||
" <td>1.0</td>\n",
|
||||
" <td>0.831188</td>\n",
|
||||
" <td>0.038508</td>\n",
|
||||
" <td>-0.016273</td>\n",
|
||||
" <td>-0.491554</td>\n",
|
||||
" <td>0.557856</td>\n",
|
||||
" <td>-0.444803</td>\n",
|
||||
" <td>True</td>\n",
|
||||
" <td>10.156217</td>\n",
|
||||
" <td>6.544583</td>\n",
|
||||
" </tr>\n",
|
||||
" </tbody>\n",
|
||||
"</table>\n",
|
||||
|
|
@ -221,36 +221,36 @@
|
|||
"</div>"
|
||||
],
|
||||
"text/plain": [
|
||||
" Z0 Z1 W0 W1 W2 W3 W4 v0 \\\n",
|
||||
"0 1.0 0.977197 -0.513249 -1.085495 -0.947936 -0.836739 -0.241959 True \n",
|
||||
"1 1.0 0.133648 1.718109 1.532397 -0.441355 0.338025 1.321070 True \n",
|
||||
"2 0.0 0.261689 2.167228 -0.133905 -0.641217 1.669584 0.555210 True \n",
|
||||
"3 1.0 0.699376 -0.164822 1.277675 -1.814968 -0.757798 0.253259 True \n",
|
||||
"4 1.0 0.669321 0.703854 0.818505 0.106033 0.139291 -0.996312 True \n",
|
||||
"... ... ... ... ... ... ... ... ... \n",
|
||||
"9995 1.0 0.952999 0.983547 1.205418 -0.845736 -1.665635 -0.285645 True \n",
|
||||
"9996 1.0 0.893868 0.147810 0.320192 -0.776280 -0.458415 -0.824306 True \n",
|
||||
"9997 0.0 0.818243 0.422534 -0.360710 -0.761213 -1.156205 0.996336 True \n",
|
||||
"9998 1.0 0.485403 0.665532 1.046824 -1.026499 0.815819 2.431939 True \n",
|
||||
"9999 0.0 0.334394 1.608105 -1.329177 -2.369921 0.185035 -0.092841 True \n",
|
||||
" Z0 Z1 W0 W1 W2 W3 W4 v0 \\\n",
|
||||
"0 1.0 0.297495 1.104464 -1.243507 2.078729 -0.035286 -1.763171 True \n",
|
||||
"1 1.0 0.940643 1.552100 0.374128 -0.198048 -0.202730 0.024421 True \n",
|
||||
"2 1.0 0.782048 0.413752 -0.404527 1.963213 1.579329 0.539134 True \n",
|
||||
"3 1.0 0.542946 -2.670037 0.290079 -0.104658 -1.642732 3.386802 False \n",
|
||||
"4 0.0 0.369742 -0.503177 0.160772 2.177942 -0.401280 1.727678 True \n",
|
||||
"... ... ... ... ... ... ... ... ... \n",
|
||||
"9995 0.0 0.314375 0.410457 0.025014 -1.307578 0.495262 0.247480 False \n",
|
||||
"9996 0.0 0.134332 -0.249336 0.936425 0.620813 1.940425 0.880728 True \n",
|
||||
"9997 0.0 0.121186 0.285559 0.554069 0.496691 -0.324387 -2.102712 True \n",
|
||||
"9998 0.0 0.400960 -0.326070 1.179469 1.196163 -0.216770 -0.173123 True \n",
|
||||
"9999 1.0 0.831188 0.038508 -0.016273 -0.491554 0.557856 -0.444803 True \n",
|
||||
"\n",
|
||||
" y \n",
|
||||
"0 -0.443120 \n",
|
||||
"1 23.899661 \n",
|
||||
"2 24.951954 \n",
|
||||
"3 8.803943 \n",
|
||||
"4 15.721275 \n",
|
||||
"0 11.701762 \n",
|
||||
"1 13.893867 \n",
|
||||
"2 22.024896 \n",
|
||||
"3 6.253032 \n",
|
||||
"4 25.160176 \n",
|
||||
"... ... \n",
|
||||
"9995 9.965975 \n",
|
||||
"9996 8.452899 \n",
|
||||
"9997 5.366551 \n",
|
||||
"9998 20.121092 \n",
|
||||
"9999 10.156217 \n",
|
||||
"9995 -2.947694 \n",
|
||||
"9996 18.983531 \n",
|
||||
"9997 4.754810 \n",
|
||||
"9998 15.567185 \n",
|
||||
"9999 6.544583 \n",
|
||||
"\n",
|
||||
"[10000 rows x 9 columns]"
|
||||
]
|
||||
},
|
||||
"execution_count": 26,
|
||||
"execution_count": 3,
|
||||
"metadata": {},
|
||||
"output_type": "execute_result"
|
||||
}
|
||||
|
|
@ -290,7 +290,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 27,
|
||||
"execution_count": 4,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -315,7 +315,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 28,
|
||||
"execution_count": 5,
|
||||
"metadata": {},
|
||||
"outputs": [],
|
||||
"source": [
|
||||
|
|
@ -324,7 +324,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 29,
|
||||
"execution_count": 6,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -352,14 +352,14 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 30,
|
||||
"execution_count": 7,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_identifier:Common causes of treatment and outcome:['W3', 'W1', 'W4', 'Unobserved Confounders', 'W0', 'W2']\n",
|
||||
"INFO:dowhy.causal_identifier:Common causes of treatment and outcome:['W2', 'W1', 'W3', 'W0', 'W4', 'Unobserved Confounders']\n",
|
||||
"WARNING:dowhy.causal_identifier:If this is observed data (not from a randomized experiment), there might always be missing confounders. Causal effect cannot be identified perfectly.\n"
|
||||
]
|
||||
},
|
||||
|
|
@ -374,27 +374,27 @@
|
|||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_identifier:Instrumental variables for treatment and outcome:['Z1', 'Z0']\n"
|
||||
"INFO:dowhy.causal_identifier:Instrumental variables for treatment and outcome:['Z0', 'Z1']\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"Estimand type: ate\n",
|
||||
"Estimand type: nonparametric-ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
" d \n",
|
||||
"─────(Expectation(y|W2,W1,W3,W0,W4))\n",
|
||||
"d[v₀] \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→{v0} and U→y then P(y|v0,W2,W1,W3,W0,W4,U) = P(y|v0,W2,W1,W3,W0,W4)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Expectation(Derivative(y, Z0)*Derivative([v0], Z0)**(-1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z0,Z1)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z0,Z1}→{v0}, then ¬({Z0,Z1}→y)\n",
|
||||
"\n"
|
||||
]
|
||||
}
|
||||
|
|
@ -415,7 +415,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 31,
|
||||
"execution_count": 8,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -423,7 +423,7 @@
|
|||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Linear Regression Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W3+W1+W4+W0+W2\n"
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W2+W1+W3+W0+W4\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
|
|
@ -433,30 +433,30 @@
|
|||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"Estimand type: nonparametric-ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
" d \n",
|
||||
"─────(Expectation(y|W2,W1,W3,W0,W4))\n",
|
||||
"d[v₀] \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→{v0} and U→y then P(y|v0,W2,W1,W3,W0,W4,U) = P(y|v0,W2,W1,W3,W0,W4)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Expectation(Derivative(y, Z0)*Derivative([v0], Z0)**(-1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z0,Z1)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z0,Z1}→{v0}, then ¬({Z0,Z1}→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"b: y~v0+W2+W1+W3+W0+W4\n",
|
||||
"## Estimate\n",
|
||||
"Value: 10.000000000000504\n",
|
||||
"Value: 9.999999999999938\n",
|
||||
"\n",
|
||||
"## Statistical Significance\n",
|
||||
"p-value: <0.001\n",
|
||||
"\n",
|
||||
"Causal Estimate is 10.000000000000504\n"
|
||||
"Causal Estimate is 9.999999999999938\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
|
|
@ -479,7 +479,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 32,
|
||||
"execution_count": 9,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -487,7 +487,7 @@
|
|||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Propensity Score Stratification Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W2+W1+W3+W0+W4\n",
|
||||
"/home/amit/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:744: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
|
||||
" y = column_or_1d(y, warn=True)\n"
|
||||
]
|
||||
|
|
@ -499,27 +499,27 @@
|
|||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"Estimand type: nonparametric-ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
" d \n",
|
||||
"─────(Expectation(y|W2,W1,W3,W0,W4))\n",
|
||||
"d[v₀] \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→{v0} and U→y then P(y|v0,W2,W1,W3,W0,W4,U) = P(y|v0,W2,W1,W3,W0,W4)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Expectation(Derivative(y, Z0)*Derivative([v0], Z0)**(-1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z0,Z1)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z0,Z1}→{v0}, then ¬({Z0,Z1}→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"b: y~v0+W2+W1+W3+W0+W4\n",
|
||||
"## Estimate\n",
|
||||
"Value: 9.945007020597561\n",
|
||||
"Value: 9.93104016350243\n",
|
||||
"\n",
|
||||
"Causal Estimate is 9.945007020597561\n"
|
||||
"Causal Estimate is 9.93104016350243\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
|
|
@ -542,7 +542,7 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 33,
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
|
|
@ -550,7 +550,7 @@
|
|||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Propensity Score Matching Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W2+W1+W3+W0+W4\n",
|
||||
"/home/amit/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:744: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
|
||||
" y = column_or_1d(y, warn=True)\n",
|
||||
"/mnt/c/Users/amshar/code/dowhy/dowhy/causal_estimators/propensity_score_matching_estimator.py:62: FutureWarning: `item` has been deprecated and will be removed in a future version\n",
|
||||
|
|
@ -558,36 +558,6 @@
|
|||
"/mnt/c/Users/amshar/code/dowhy/dowhy/causal_estimators/propensity_score_matching_estimator.py:77: FutureWarning: `item` has been deprecated and will be removed in a future version\n",
|
||||
" treated_outcome = treated.iloc[indices[i]][self._outcome_name].item()\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"## Estimate\n",
|
||||
"Value: 10.106764870393203\n",
|
||||
"\n",
|
||||
"Causal Estimate is 10.106764870393203\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"source": [
|
||||
|
|
@ -612,56 +582,9 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 38,
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Propensity Score Weighting Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:b: y~v0+W3+W1+W4+W0+W2\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"b: y~v0+W3+W1+W4+W0+W2\n",
|
||||
"## Estimate\n",
|
||||
"Value: 11.152230663995484\n",
|
||||
"\n",
|
||||
"Causal Estimate is 11.152230663995484\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"/home/amit/.local/lib/python3.6/site-packages/sklearn/utils/validation.py:744: DataConversionWarning: A column-vector y was passed when a 1d array was expected. Please change the shape of y to (n_samples, ), for example using ravel().\n",
|
||||
" y = column_or_1d(y, warn=True)\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"causal_estimate_ipw = model.estimate_effect(identified_estimand,\n",
|
||||
" method_name=\"backdoor.propensity_score_weighting\",\n",
|
||||
|
|
@ -682,68 +605,11 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 35,
|
||||
"execution_count": null,
|
||||
"metadata": {
|
||||
"scrolled": true
|
||||
},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Instrumental Variable Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:Realized estimand: Wald Estimator\n",
|
||||
"Realized estimand type: ate\n",
|
||||
"Estimand expression:\n",
|
||||
" -1\n",
|
||||
"Expectation(Derivative(y, Z1))⋅Expectation(Derivative(v0, Z1)) \n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Estimand assumption 3, treatment_effect_homogeneity: Each unit's treatment ['v0'] isaffected in the same way by common causes of ['v0'] and y\n",
|
||||
"Estimand assumption 4, outcome_effect_homogeneity: Each unit's outcome y isaffected in the same way by common causes of ['v0'] and y\n",
|
||||
"\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"Realized estimand: Wald Estimator\n",
|
||||
"Realized estimand type: ate\n",
|
||||
"Estimand expression:\n",
|
||||
" -1\n",
|
||||
"Expectation(Derivative(y, Z1))⋅Expectation(Derivative(v0, Z1)) \n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Estimand assumption 3, treatment_effect_homogeneity: Each unit's treatment ['v0'] isaffected in the same way by common causes of ['v0'] and y\n",
|
||||
"Estimand assumption 4, outcome_effect_homogeneity: Each unit's outcome y isaffected in the same way by common causes of ['v0'] and y\n",
|
||||
"\n",
|
||||
"## Estimate\n",
|
||||
"Value: 9.082237126676432\n",
|
||||
"\n",
|
||||
"Causal Estimate is 9.082237126676432\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"causal_estimate_iv = model.estimate_effect(identified_estimand,\n",
|
||||
" method_name=\"iv.instrumental_variable\", method_params = {'iv_instrument_name': 'Z0'})\n",
|
||||
|
|
@ -762,82 +628,9 @@
|
|||
},
|
||||
{
|
||||
"cell_type": "code",
|
||||
"execution_count": 36,
|
||||
"execution_count": null,
|
||||
"metadata": {},
|
||||
"outputs": [
|
||||
{
|
||||
"name": "stderr",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
"INFO:dowhy.causal_estimator:Using Regression Discontinuity Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:\n",
|
||||
"INFO:dowhy.causal_estimator:INFO: Using Instrumental Variable Estimator\n",
|
||||
"INFO:dowhy.causal_estimator:Realized estimand: Wald Estimator\n",
|
||||
"Realized estimand type: ate\n",
|
||||
"Estimand expression:\n",
|
||||
" -1\n",
|
||||
"Expectation(Derivative(y, Z1))⋅Expectation(Derivative(v0, Z1)) \n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Estimand assumption 3, treatment_effect_homogeneity: Each unit's treatment ['local_treatment'] isaffected in the same way by common causes of ['local_treatment'] and local_outcome\n",
|
||||
"Estimand assumption 4, outcome_effect_homogeneity: Each unit's outcome local_outcome isaffected in the same way by common causes of ['local_treatment'] and local_outcome\n",
|
||||
"\n"
|
||||
]
|
||||
},
|
||||
{
|
||||
"name": "stdout",
|
||||
"output_type": "stream",
|
||||
"text": [
|
||||
" local_rd_variable local_treatment local_outcome\n",
|
||||
"22 0.413325 True 1.597864\n",
|
||||
"31 0.567281 True 11.847969\n",
|
||||
"33 0.558196 False -1.704836\n",
|
||||
"35 0.589582 True 15.232575\n",
|
||||
"37 0.415546 True 12.738279\n",
|
||||
"... ... ... ...\n",
|
||||
"9981 0.503160 True 10.459732\n",
|
||||
"9986 0.405848 True 10.421814\n",
|
||||
"9989 0.404516 True 14.520292\n",
|
||||
"9990 0.551019 True 19.378382\n",
|
||||
"9998 0.485403 True 20.121092\n",
|
||||
"\n",
|
||||
"[1970 rows x 3 columns]\n",
|
||||
"*** Causal Estimate ***\n",
|
||||
"\n",
|
||||
"## Target estimand\n",
|
||||
"Estimand type: ate\n",
|
||||
"### Estimand : 1\n",
|
||||
"Estimand name: backdoor\n",
|
||||
"Estimand expression:\n",
|
||||
" d \n",
|
||||
"───(Expectation(y|W3,W1,W4,W0,W2))\n",
|
||||
"dv₀ \n",
|
||||
"Estimand assumption 1, Unconfoundedness: If U→v0 and U→y then P(y|v0,W3,W1,W4,W0,W2,U) = P(y|v0,W3,W1,W4,W0,W2)\n",
|
||||
"### Estimand : 2\n",
|
||||
"Estimand name: iv\n",
|
||||
"Estimand expression:\n",
|
||||
"Expectation(Derivative(y, Z1)/Derivative(v0, Z1))\n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"\n",
|
||||
"## Realized estimand\n",
|
||||
"Realized estimand: Wald Estimator\n",
|
||||
"Realized estimand type: ate\n",
|
||||
"Estimand expression:\n",
|
||||
" -1\n",
|
||||
"Expectation(Derivative(y, Z1))⋅Expectation(Derivative(v0, Z1)) \n",
|
||||
"Estimand assumption 1, As-if-random: If U→→y then ¬(U →→Z1,Z0)\n",
|
||||
"Estimand assumption 2, Exclusion: If we remove {Z1,Z0}→v0, then ¬(Z1,Z0→y)\n",
|
||||
"Estimand assumption 3, treatment_effect_homogeneity: Each unit's treatment ['local_treatment'] isaffected in the same way by common causes of ['local_treatment'] and local_outcome\n",
|
||||
"Estimand assumption 4, outcome_effect_homogeneity: Each unit's outcome local_outcome isaffected in the same way by common causes of ['local_treatment'] and local_outcome\n",
|
||||
"\n",
|
||||
"## Estimate\n",
|
||||
"Value: 10.354383271697392\n",
|
||||
"\n",
|
||||
"Causal Estimate is 10.354383271697392\n"
|
||||
]
|
||||
}
|
||||
],
|
||||
"outputs": [],
|
||||
"source": [
|
||||
"causal_estimate_regdist = model.estimate_effect(identified_estimand,\n",
|
||||
" method_name=\"iv.regression_discontinuity\", \n",
|
||||
|
|
|
|||
Различия файлов скрыты, потому что одна или несколько строк слишком длинны
|
|
@ -71,6 +71,7 @@ class CausalGraph:
|
|||
if missing_nodes_as_confounders:
|
||||
self._graph = self.add_missing_nodes_as_common_causes(observed_node_names)
|
||||
self._graph = self.add_node_attributes(observed_node_names)
|
||||
#TODO do not add it here. CausalIdentifier should call causal_graph to add an unobserved common cause if needed. This also ensures that we do not need get_common_causes in this class.
|
||||
self._graph = self.add_unobserved_common_cause(observed_node_names)
|
||||
|
||||
def view_graph(self, layout="dot"):
|
||||
|
|
@ -220,6 +221,7 @@ class CausalGraph:
|
|||
"""
|
||||
Assume that nodes1 causes nodes2 (e.g., nodes1 are the treatments and nodes2 are the outcomes)
|
||||
"""
|
||||
# TODO Refactor to remove this from here and only implement this logic in causalIdentifier. Unnecessary assumption of nodes1 to be causing nodes2.
|
||||
nodes1 = parse_state(nodes1)
|
||||
nodes2 = parse_state(nodes2)
|
||||
causes_1 = set()
|
||||
|
|
|
|||
|
|
@ -79,29 +79,32 @@ class CausalIdentifier:
|
|||
outcome_name, common_causes):
|
||||
# TODO: outputs string for now, but ideally should do symbolic
|
||||
# expressions Mon 19 Feb 2018 04:54:17 PM DST
|
||||
# TODO Better support for multivariate treatments
|
||||
|
||||
expr = None
|
||||
if estimand_type == "ate":
|
||||
# [TODO: support multivariate states]
|
||||
if estimand_type == "nonparametric-ate":
|
||||
outcome_name = outcome_name[0]
|
||||
treatment_name = treatment_name[0]
|
||||
num_expr_str = outcome_name
|
||||
if len(common_causes)>0:
|
||||
num_expr_str += "|" + ",".join(common_causes)
|
||||
expr = "d(" + num_expr_str + ")/d" + treatment_name
|
||||
expr = "d(" + num_expr_str + ")/d" + ",".join(treatment_name)
|
||||
sym_mu = sp.Symbol("mu")
|
||||
sym_sigma = sp.Symbol("sigma", positive=True)
|
||||
sym_outcome = spstats.Normal(num_expr_str, sym_mu, sym_sigma)
|
||||
# sym_common_causes = [sp.stats.Normal(common_cause, sym_mu, sym_sigma) for common_cause in common_causes]
|
||||
sym_treatment = sp.Symbol(treatment_name)
|
||||
sym_treatment_symbols = [sp.Symbol(t) for t in treatment_name]
|
||||
sym_treatment = sp.Array(sym_treatment_symbols)
|
||||
sym_conditional_outcome = spstats.Expectation(sym_outcome)
|
||||
sym_effect = sp.Derivative(sym_conditional_outcome, sym_treatment)
|
||||
|
||||
sym_assumptions = {
|
||||
'Unconfoundedness': (
|
||||
u"If U\N{RIGHTWARDS ARROW}{0} and U\N{RIGHTWARDS ARROW}{1}"
|
||||
u"If U\N{RIGHTWARDS ARROW}{{{0}}} and U\N{RIGHTWARDS ARROW}{1}"
|
||||
" then P({1}|{0},{2},U) = P({1}|{0},{2})"
|
||||
).format(treatment_name, outcome_name, ",".join(common_causes))
|
||||
).format(",".join(treatment_name), outcome_name, ",".join(common_causes))
|
||||
}
|
||||
else:
|
||||
raise ValueError("Estimand type not supported. Supported estimand types are 'non-parametric-ate'.")
|
||||
|
||||
estimand = {
|
||||
'estimand': sym_effect,
|
||||
|
|
@ -111,13 +114,14 @@ class CausalIdentifier:
|
|||
|
||||
def construct_iv_estimand(self, estimand_type, treatment_name,
|
||||
outcome_name, instrument_names):
|
||||
# TODO: support multivariate treatments better.
|
||||
expr = None
|
||||
if estimand_type == "ate":
|
||||
# [TODO: support multivariate states]
|
||||
if estimand_type == "nonparametric-ate":
|
||||
outcome_name = outcome_name[0]
|
||||
treatment_name = treatment_name[0]
|
||||
sym_outcome = spstats.Normal(outcome_name, 0, 1)
|
||||
sym_treatment = spstats.Normal(treatment_name, 0, 1)
|
||||
sym_treatment_symbols = [spstats.Normal(t, 0, 1) for t in treatment_name]
|
||||
sym_treatment = sp.Array(sym_treatment_symbols)
|
||||
#sym_treatment = spstats.Normal(treatment_name, 0, 1)
|
||||
sym_instrument = sp.Symbol(instrument_names[0]) # ",".join(instrument_names))
|
||||
sym_outcome_derivative = sp.Derivative(sym_outcome, sym_instrument)
|
||||
sym_treatment_derivative = sp.Derivative(sym_treatment, sym_instrument)
|
||||
|
|
@ -128,11 +132,13 @@ class CausalIdentifier:
|
|||
"\N{NOT SIGN}(U \N{RIGHTWARDS ARROW}\N{RIGHTWARDS ARROW}{1})"
|
||||
).format(outcome_name, ",".join(instrument_names)),
|
||||
"Exclusion": (
|
||||
u"If we remove {{{0}}}\N{RIGHTWARDS ARROW}{1}, then "
|
||||
u"\N{NOT SIGN}({0}\N{RIGHTWARDS ARROW}{2})"
|
||||
).format(",".join(instrument_names), treatment_name,
|
||||
u"If we remove {{{0}}}\N{RIGHTWARDS ARROW}{{{1}}}, then "
|
||||
u"\N{NOT SIGN}({{{0}}}\N{RIGHTWARDS ARROW}{2})"
|
||||
).format(",".join(instrument_names), ",".join(treatment_name),
|
||||
outcome_name)
|
||||
}
|
||||
else:
|
||||
raise ValueError("Estimand type not supported. Supported estimand types are 'non-parametric-ate'.")
|
||||
|
||||
estimand = {
|
||||
'estimand': sym_effect,
|
||||
|
|
|
|||
|
|
@ -27,7 +27,7 @@ class CausalModel:
|
|||
def __init__(self, data, treatment, outcome, graph=None,
|
||||
common_causes=None, instruments=None,
|
||||
effect_modifiers=None,
|
||||
estimand_type="ate",
|
||||
estimand_type="nonparametric-ate",
|
||||
proceed_when_unidentifiable=False,
|
||||
missing_nodes_as_confounders=False,
|
||||
**kwargs):
|
||||
|
|
@ -49,7 +49,7 @@ class CausalModel:
|
|||
:param instruments: names of instrumental variables for the effect of
|
||||
treatment on outcome
|
||||
:param effect_modifiers: names of variables that can modify the treatment effect (useful for heterogeneous treatment effect estimation)
|
||||
:param estimand_type: the type of estimand requested (can be "ate" for average treatment effect, "att" for average treatment effect on the treated", and "atc", for average treatment effect on the control population.
|
||||
:param estimand_type: the type of estimand requested (currently only "nonparametric-ate" is supported). In the future, may support other specific parametric forms of identification.
|
||||
:proceed_when_unidentifiable: does the identification proceed by ignoring potential unobserved confounders. Binary flag.
|
||||
:missing_nodes_as_confounders: Binary flag indicating whether variables in the dataframe that are not included in the causal graph, should be automatically included as confounder nodes.
|
||||
|
||||
|
|
|
|||
Загрузка…
Ссылка в новой задаче