{ "cells": [ { "cell_type": "markdown", "metadata": {}, "source": [ "# PyCaret 2 Regression Example\n", "This notebook is created using PyCaret 2.0. Last updated : 31-07-2020" ] }, { "cell_type": "code", "execution_count": 1, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "2.0\n" ] } ], "source": [ "# check version\n", "from pycaret.utils import version\n", "version()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 1. Loading Dataset" ] }, { "cell_type": "code", "execution_count": 2, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
agesexbmichildrensmokerregioncharges
019female27.9000yessouthwest16884.92400
118male33.7701nosoutheast1725.55230
228male33.0003nosoutheast4449.46200
333male22.7050nonorthwest21984.47061
432male28.8800nonorthwest3866.85520
\n", "
" ], "text/plain": [ " age sex bmi children smoker region charges\n", "0 19 female 27.900 0 yes southwest 16884.92400\n", "1 18 male 33.770 1 no southeast 1725.55230\n", "2 28 male 33.000 3 no southeast 4449.46200\n", "3 33 male 22.705 0 no northwest 21984.47061\n", "4 32 male 28.880 0 no northwest 3866.85520" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "from pycaret.datasets import get_data\n", "data = get_data('insurance')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 2. Initialize Setup" ] }, { "cell_type": "code", "execution_count": 3, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ " \n", "Setup Succesfully Completed.\n" ] }, { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Description Value
0session_id123
1Transform Target False
2Transform Target MethodNone
3Original Data(1338, 7)
4Missing Values False
5Numeric Features 2
6Categorical Features 4
7Ordinal Features False
8High Cardinality Features False
9High Cardinality Method None
10Sampled Data(1338, 7)
11Transformed Train Set(936, 16)
12Transformed Test Set(402, 16)
13Numeric Imputer mean
14Categorical Imputer constant
15Normalize False
16Normalize Method None
17Transformation False
18Transformation Method None
19PCA False
20PCA Method None
21PCA Components None
22Ignore Low Variance False
23Combine Rare Levels False
24Rare Level Threshold None
25Numeric Binning False
26Remove Outliers False
27Outliers Threshold None
28Remove Multicollinearity False
29Multicollinearity Threshold None
30Clustering False
31Clustering Iteration None
32Polynomial Features False
33Polynomial Degree None
34Trignometry Features False
35Polynomial Threshold None
36Group Features False
37Feature Selection False
38Features Selection Threshold None
39Feature Interaction False
40Feature Ratio False
41Interaction Threshold None
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "name": "stderr", "output_type": "stream", "text": [ "2020/07/31 08:44:23 WARNING mlflow.tracking.context.git_context: Failed to import Git (the Git executable is probably not on your PATH), so Git SHA is not available. Error: No module named 'repository'\n" ] } ], "source": [ "from pycaret.regression import *\n", "reg1 = setup(data, target = 'charges', session_id=123, log_experiment=True, experiment_name='insurance1')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 3. Compare Baseline" ] }, { "cell_type": "code", "execution_count": 4, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
Model MAE MSE RMSE R2 RMSLE MAPE TT (Sec)
0Extreme Gradient Boosting2666.867522721899.53054764.42280.84100.44280.31510.0429
1Gradient Boosting Regressor2671.592723019681.26614794.60370.83930.44390.31430.0683
2CatBoost Regressor2814.604824757340.46594973.77650.82650.47340.34271.1286
3Random Forest2779.202625351757.15065032.25870.82180.48160.34320.2087
4Light Gradient Boosting Machine3018.989525515012.30515049.84920.81920.55340.38760.0815
5Extra Trees Regressor2755.926528180447.26585299.65660.80430.48750.32550.1496
6AdaBoost Regressor4366.100129298215.00875411.06060.79150.64780.76620.0195
7Ridge Regression4339.609338542499.62026196.48910.73430.63480.44290.0036
8Bayesian Ridge4343.500638542310.25366196.46070.73430.64050.44360.0058
9Linear Regression4332.765838549952.00266197.08420.73430.63690.44150.0043
10Lasso Regression4332.632738543897.46926196.60740.73430.64040.44160.0038
11TheilSen Regressor4124.365838946435.26316224.89170.73270.53370.37430.7617
12Least Angle Regression4323.457840017870.22866312.01150.72500.56470.42420.0062
13Lasso Least Angle Regression4322.446640023599.45506312.54980.72490.54010.42450.0060
14Decision Tree3184.972844561182.45696663.22480.68260.53430.35230.0050
15Huber Regressor3478.863549170605.58596997.82280.65900.48730.22120.0472
16Random Sample Consensus3467.403652056856.30747203.07740.63820.49700.21750.0831
17Orthogonal Matching Pursuit5760.047557656797.20767580.02240.60260.74260.89960.0034
18Passive Aggressive Regressor4817.272659211213.83237652.97070.59280.75770.43340.0068
19Elastic Net6399.470272811792.65778506.28130.50210.67890.80160.0030
20K Neighbors Regressor6858.1227105272520.336310228.24970.27840.75240.74500.0036
21Support Vector Machine8401.1273163965107.005212732.6249-0.10920.93031.03230.0308
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "best_model = compare_models(fold=5)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 4. Create Model" ] }, { "cell_type": "code", "execution_count": 5, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
02972.932924058897.43484904.98700.85630.60830.3985
13080.534029299758.44805412.92510.82960.44590.3268
23022.231427624562.66355255.90740.65020.68220.4361
33146.242225018958.48765001.89550.79110.63980.5095
43154.689928894513.08805375.36170.78510.59010.3617
52931.089621432486.19794629.52330.86210.41310.2829
62625.935820785814.44154559.14620.85860.38910.3070
72678.361724232738.55404922.67600.86680.50980.2720
82710.338021418665.48184628.03040.86170.55740.4101
93273.228629598375.85945440.43890.80940.59160.3804
Mean2959.558425236477.06565013.08920.81710.54270.3685
SD210.55993254481.9214324.67550.06280.09430.0703
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "lightgbm = create_model('lightgbm')" ] }, { "cell_type": "code", "execution_count": 6, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
04202.083037315426.00496108.63540.77710.69690.5901
14164.762238772693.89766226.77230.77460.68260.4917
24545.748945041923.56546711.32800.42970.95580.6772
34390.337041942311.13626476.28840.64980.71400.7225
44617.241741981223.15756479.29190.68780.69630.5784
54166.022534525507.14055875.84100.77790.61990.4641
63916.845631704830.13375630.70420.78430.71890.4801
73689.608832718649.01735720.02180.82010.62660.4303
83995.668535808315.21885984.00490.76870.90640.6267
94616.177247556279.27636896.10610.69370.86570.7124
Mean4230.449538736715.85486210.89940.71640.74830.5773
SD296.41535023332.8140401.80160.10810.11170.1015
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "import numpy as np\n", "lgbms = [create_model('lightgbm', learning_rate=i) for i in np.arange(0.1,1,0.1)]" ] }, { "cell_type": "code", "execution_count": 8, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "9\n" ] } ], "source": [ "print(len(lgbms))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 5. Tune Hyperparameters" ] }, { "cell_type": "code", "execution_count": 9, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
02593.520318901959.35124347.63840.88710.41220.3237
13035.882630123714.30975488.50750.82490.45770.3318
22783.871021407816.74634626.85820.72900.49230.4139
32870.694521243674.28784609.08610.82260.47480.4161
42838.747324370286.38644936.62700.81870.44550.3016
52632.637918784222.60474334.07690.87920.35880.2780
62523.077419451327.41764410.36590.86760.38880.3309
72700.145924627482.80874962.60850.86460.46740.3220
82627.214219952600.02444466.83330.87110.46010.3909
92945.212525414198.63015041.24970.83630.49710.3744
Mean2755.100422427728.25674722.38510.84010.44550.3483
SD158.26643462560.6654356.09940.04410.04290.0452
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "tuned_lightgbm = tune_model(lightgbm, n_iter=50, optimize = 'MAE')" ] }, { "cell_type": "code", "execution_count": 10, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "LGBMRegressor(boosting_type='gbdt', class_weight=None, colsample_bytree=1.0,\n", " importance_type='split', learning_rate=0.3, max_depth=70,\n", " min_child_samples=20, min_child_weight=0.001, min_split_gain=0.2,\n", " n_estimators=10, n_jobs=-1, num_leaves=10, objective=None,\n", " random_state=123, reg_alpha=0.4, reg_lambda=0.1, silent=True,\n", " subsample=1.0, subsample_for_bin=200000, subsample_freq=0)" ] }, "execution_count": 10, "metadata": {}, "output_type": "execute_result" } ], "source": [ "tuned_lightgbm" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 6. Ensemble Model" ] }, { "cell_type": "code", "execution_count": 11, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
03130.752839833455.18046311.37510.76210.50670.3295
13104.836945066079.66636713.12740.73800.54100.3104
23315.342047502769.51276892.22530.39860.64520.4540
32869.007840057346.77296329.08740.66550.60430.5635
44039.388164499878.17568031.18160.52030.65900.5106
53324.741941203145.55546418.96760.73500.49160.3224
62579.693338790767.70846228.22350.73610.39110.2259
72727.953035755628.99505979.60110.80340.46980.1807
82863.010638662493.23256217.91710.75030.51130.4390
93207.843447265363.54626874.98100.69560.51210.2716
Mean3116.257043863692.83456599.66870.68050.53320.3608
SD388.32297800317.0820555.03680.11900.07810.1191
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "dt = create_model('dt')" ] }, { "cell_type": "code", "execution_count": 12, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
02689.426422734130.61234768.03220.86420.47560.3477
12850.548930834672.86075552.89770.82070.46830.2826
22767.049924433673.62384943.04300.69060.54590.3887
32842.443824548531.64394954.64750.79500.51650.4156
43020.600230696946.66765540.48250.77170.56020.3738
52818.944222660137.14364760.26650.85420.37120.2631
62617.322022836756.39734778.78190.84460.38740.3035
72684.810124880599.57404988.04570.86320.44510.2565
82334.316118535034.55314305.23340.88030.43570.3510
92820.224929299167.65515412.87060.81130.50720.3478
Mean2744.568625145965.07315000.43010.81960.47130.3330
SD173.16943773482.7721376.38290.05380.05990.0515
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "bagged_dt = ensemble_model(dt, n_estimators=50)" ] }, { "cell_type": "code", "execution_count": 13, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
02088.609825306132.77245030.52010.84890.43600.2089
12694.450037713226.91656141.10960.78070.50230.2522
22580.290431829320.07265641.74800.59700.59540.3725
32026.386722848511.31274780.01160.80920.38170.1531
42783.996936022333.05886001.86080.73210.58990.3035
53304.082042910088.53646550.57930.72400.43980.2454
61691.509121151389.59784599.06400.85610.32260.1226
71941.823122752869.91424769.99680.87490.36080.1170
81842.404420941293.38474576.16580.86470.36410.2280
92618.303336220824.96726018.37390.76670.48300.2274
Mean2357.185629769599.05335410.94300.78540.44760.2231
SD487.10357681159.5886700.92440.08150.09000.0753
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "boosted_dt = ensemble_model(dt, method = 'Boosting')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 7. Blend Models" ] }, { "cell_type": "code", "execution_count": 14, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
03567.602927065352.26045202.43710.83830.44290.3872
13948.544338042231.79246167.83850.77880.46810.3602
23248.690723594846.35484857.45270.70130.50820.4408
33228.900621265240.68744611.42500.82240.48460.4734
43810.779433931591.30725825.08290.74760.49260.3832
53393.462627950743.30645286.84630.82020.37290.3003
63321.494429360441.72895418.52760.80020.42150.3687
73924.089832840540.14065730.66660.81940.44480.3407
83590.269528114541.81155302.31480.81840.47470.4435
93792.933532627264.43485712.02800.78980.50950.3913
Mean3582.676829479279.38245411.46200.79370.46200.3889
SD262.23434759119.6736441.99420.03970.04030.0493
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "blender = blend_models()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 8. Stack Models" ] }, { "cell_type": "code", "execution_count": 15, "metadata": {}, "outputs": [ { "data": { "text/html": [ "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
MAE MSE RMSE R2 RMSLE MAPE
02564.580920266074.93454501.78570.87900.43700.3141
12934.377031081553.12675575.08320.81930.49790.2816
22670.167223415471.97124838.95360.70350.50990.3105
32798.866521176854.59814601.83170.82320.46310.3736
42894.412527549639.01335248.77500.79510.52080.3140
52684.300120037070.56934476.27870.87110.45710.2697
62458.370619953697.91064466.95620.86420.34860.2697
72744.866425598490.32655059.49510.85920.43470.2507
82341.073817775685.53444216.12210.88520.39640.3114
92994.982127582079.60675251.86440.82230.53550.3599
Mean2708.599723443661.75914823.71460.83220.46010.3055
SD199.36874100160.4270418.85500.05150.05570.0373
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "stacker = stack_models(estimator_list = compare_models(n_select=5, fold = 5, whitelist = models(type='ensemble').index.tolist()))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 9. Analyze Model" ] }, { "cell_type": "code", "execution_count": 16, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plot_model(dt)" ] }, { "cell_type": "code", "execution_count": 17, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plot_model(dt, plot = 'error')" ] }, { "cell_type": "code", "execution_count": 18, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "plot_model(dt, plot = 'feature')" ] }, { "cell_type": "code", "execution_count": 19, "metadata": {}, "outputs": [ { "data": { "application/vnd.jupyter.widget-view+json": { "model_id": "40d304e225f94054b49f71f91a7bb130", "version_major": 2, "version_minor": 0 }, "text/plain": [ "interactive(children=(ToggleButtons(description='Plot Type:', icons=('',), options=(('Hyperparameters', 'param…" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "evaluate_model(dt)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 10. Interpret Model" ] }, { "cell_type": "code", "execution_count": 20, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "interpret_model(lightgbm)" ] }, { "cell_type": "code", "execution_count": 21, "metadata": {}, "outputs": [ { "data": { "image/png": "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\n", "text/plain": [ "
" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "interpret_model(lightgbm, plot = 'correlation')" ] }, { "cell_type": "code", "execution_count": 22, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
" ], "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "\n", "
\n", "
\n", " Visualization omitted, Javascript library not loaded!
\n", " Have you run `initjs()` in this notebook? If this notebook was from another\n", " user you must also trust this notebook (File -> Trust notebook). If you are viewing\n", " this notebook on github the Javascript has been stripped for security. If you are using\n", " JupyterLab this error is because a JupyterLab extension has not yet been written.\n", "
\n", " " ], "text/plain": [ "" ] }, "execution_count": 22, "metadata": {}, "output_type": "execute_result" } ], "source": [ "interpret_model(lightgbm, plot = 'reason', observation = 12)" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 11. AutoML()" ] }, { "cell_type": "code", "execution_count": 23, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "AdaBoostRegressor(base_estimator=DecisionTreeRegressor(ccp_alpha=0.0,\n", " criterion='mse',\n", " max_depth=None,\n", " max_features=None,\n", " max_leaf_nodes=None,\n", " min_impurity_decrease=0.0,\n", " min_impurity_split=None,\n", " min_samples_leaf=1,\n", " min_samples_split=2,\n", " min_weight_fraction_leaf=0.0,\n", " presort='deprecated',\n", " random_state=123,\n", " splitter='best'),\n", " learning_rate=1.0, loss='linear', n_estimators=10,\n", " random_state=123)" ] }, "execution_count": 23, "metadata": {}, "output_type": "execute_result" } ], "source": [ "best = automl(optimize = 'MAE')\n", "best" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 12. Predict Model" ] }, { "cell_type": "code", "execution_count": 24, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
ModelMAEMSERMSER2RMSLEMAPE
0Light Gradient Boosting Machine2764.03532.068683e+074548.27730.85520.52230.3678
\n", "
" ], "text/plain": [ " Model MAE MSE RMSE \\\n", "0 Light Gradient Boosting Machine 2764.0353 2.068683e+07 4548.2773 \n", "\n", " R2 RMSLE MAPE \n", "0 0.8552 0.5223 0.3678 " ] }, "metadata": {}, "output_type": "display_data" }, { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
agebmisex_femalesex_malechildren_0children_1children_2children_3children_4children_5smoker_nosmoker_yesregion_northeastregion_northwestregion_southeastregion_southwestchargesLabel
049.042.6801.00.00.00.01.00.00.00.01.00.00.00.01.00.09800.888209363.9665
132.037.3350.01.00.01.00.00.00.00.01.00.01.00.00.00.04667.6076510393.4936
227.031.4001.00.01.00.00.00.00.00.00.01.00.00.00.01.034838.8730035628.9933
335.024.1300.01.00.01.00.00.00.00.01.00.00.01.00.00.05125.215706926.8267
460.025.7400.01.01.00.00.00.00.00.01.00.00.00.01.00.012142.5786017508.2783
\n", "
" ], "text/plain": [ " age bmi sex_female sex_male children_0 children_1 children_2 \\\n", "0 49.0 42.680 1.0 0.0 0.0 0.0 1.0 \n", "1 32.0 37.335 0.0 1.0 0.0 1.0 0.0 \n", "2 27.0 31.400 1.0 0.0 1.0 0.0 0.0 \n", "3 35.0 24.130 0.0 1.0 0.0 1.0 0.0 \n", "4 60.0 25.740 0.0 1.0 1.0 0.0 0.0 \n", "\n", " children_3 children_4 children_5 smoker_no smoker_yes \\\n", "0 0.0 0.0 0.0 1.0 0.0 \n", "1 0.0 0.0 0.0 1.0 0.0 \n", "2 0.0 0.0 0.0 0.0 1.0 \n", "3 0.0 0.0 0.0 1.0 0.0 \n", "4 0.0 0.0 0.0 1.0 0.0 \n", "\n", " region_northeast region_northwest region_southeast region_southwest \\\n", "0 0.0 0.0 1.0 0.0 \n", "1 1.0 0.0 0.0 0.0 \n", "2 0.0 0.0 0.0 1.0 \n", "3 0.0 1.0 0.0 0.0 \n", "4 0.0 0.0 1.0 0.0 \n", "\n", " charges Label \n", "0 9800.88820 9363.9665 \n", "1 4667.60765 10393.4936 \n", "2 34838.87300 35628.9933 \n", "3 5125.21570 6926.8267 \n", "4 12142.57860 17508.2783 " ] }, "execution_count": 24, "metadata": {}, "output_type": "execute_result" } ], "source": [ "pred_holdouts = predict_model(lightgbm)\n", "pred_holdouts.head()" ] }, { "cell_type": "code", "execution_count": 25, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
agesexbmichildrensmokerregionLabel
019female27.9000yessouthwest16884.9240
118male33.7701nosoutheast1725.5523
228male33.0003nosoutheast5138.2567
333male22.7050nonorthwest21984.4706
432male28.8800nonorthwest3866.8552
\n", "
" ], "text/plain": [ " age sex bmi children smoker region Label\n", "0 19 female 27.900 0 yes southwest 16884.9240\n", "1 18 male 33.770 1 no southeast 1725.5523\n", "2 28 male 33.000 3 no southeast 5138.2567\n", "3 33 male 22.705 0 no northwest 21984.4706\n", "4 32 male 28.880 0 no northwest 3866.8552" ] }, "execution_count": 25, "metadata": {}, "output_type": "execute_result" } ], "source": [ "new_data = data.copy()\n", "new_data.drop(['charges'], axis=1, inplace=True)\n", "predict_new = predict_model(best, data=new_data)\n", "predict_new.head()" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 13. Save / Load Model" ] }, { "cell_type": "code", "execution_count": 26, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Transformation Pipeline and Model Succesfully Saved\n" ] } ], "source": [ "save_model(best, model_name='best-model')" ] }, { "cell_type": "code", "execution_count": 27, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Transformation Pipeline and Model Sucessfully Loaded\n", "[Pipeline(memory=None,\n", " steps=[('dtypes',\n", " DataTypes_Auto_infer(categorical_features=[],\n", " display_types=True, features_todrop=[],\n", " ml_usecase='regression',\n", " numerical_features=[], target='charges',\n", " time_features=[])),\n", " ('imputer',\n", " Simple_Imputer(categorical_strategy='not_available',\n", " numeric_strategy='mean',\n", " target_variable=None)),\n", " ('new_levels1',\n", " New_Catagorical_Levels...\n", " ('group', Empty()), ('nonliner', Empty()), ('scaling', Empty()),\n", " ('P_transform', Empty()), ('pt_target', Empty()),\n", " ('binn', Empty()), ('rem_outliers', Empty()),\n", " ('cluster_all', Empty()), ('dummy', Dummify(target='charges')),\n", " ('fix_perfect', Empty()), ('clean_names', Clean_Colum_Names()),\n", " ('feature_select', Empty()), ('fix_multi', Empty()),\n", " ('dfs', Empty()), ('pca', Empty())],\n", " verbose=False), AdaBoostRegressor(base_estimator=DecisionTreeRegressor(ccp_alpha=0.0,\n", " criterion='mse',\n", " max_depth=None,\n", " max_features=None,\n", " max_leaf_nodes=None,\n", " min_impurity_decrease=0.0,\n", " min_impurity_split=None,\n", " min_samples_leaf=1,\n", " min_samples_split=2,\n", " min_weight_fraction_leaf=0.0,\n", " presort='deprecated',\n", " random_state=123,\n", " splitter='best'),\n", " learning_rate=1.0, loss='linear', n_estimators=10,\n", " random_state=123), None]\n" ] } ], "source": [ "loaded_bestmodel = load_model('best-model')\n", "print(loaded_bestmodel)" ] }, { "cell_type": "code", "execution_count": 28, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
Pipeline(memory=None,\n",
       "         steps=[('dtypes',\n",
       "                 DataTypes_Auto_infer(categorical_features=[],\n",
       "                                      display_types=True, features_todrop=[],\n",
       "                                      ml_usecase='regression',\n",
       "                                      numerical_features=[], target='charges',\n",
       "                                      time_features=[])),\n",
       "                ('imputer',\n",
       "                 Simple_Imputer(categorical_strategy='not_available',\n",
       "                                numeric_strategy='mean',\n",
       "                                target_variable=None)),\n",
       "                ('new_levels1',\n",
       "                 New_Catagorical_Levels...\n",
       "                ('group', Empty()), ('nonliner', Empty()), ('scaling', Empty()),\n",
       "                ('P_transform', Empty()), ('pt_target', Empty()),\n",
       "                ('binn', Empty()), ('rem_outliers', Empty()),\n",
       "                ('cluster_all', Empty()), ('dummy', Dummify(target='charges')),\n",
       "                ('fix_perfect', Empty()), ('clean_names', Clean_Colum_Names()),\n",
       "                ('feature_select', Empty()), ('fix_multi', Empty()),\n",
       "                ('dfs', Empty()), ('pca', Empty())],\n",
       "         verbose=False)
DataTypes_Auto_infer(ml_usecase='regression', target='charges')
Simple_Imputer(categorical_strategy='not_available', numeric_strategy='mean',\n",
       "               target_variable=None)
New_Catagorical_Levels_in_TestData(replacement_strategy='least frequent',\n",
       "                                   target='charges')
Empty()
Empty()
Empty()
Empty()
New_Catagorical_Levels_in_TestData(replacement_strategy='least frequent',\n",
       "                                   target='charges')
Make_Time_Features(list_of_features=None)
Empty()
Empty()
Empty()
Empty()
Empty()
Empty()
Empty()
Empty()
Dummify(target='charges')
Empty()
Clean_Colum_Names()
Empty()
Empty()
Empty()
Empty()
" ], "text/plain": [ "Pipeline(memory=None,\n", " steps=[('dtypes',\n", " DataTypes_Auto_infer(categorical_features=[],\n", " display_types=True, features_todrop=[],\n", " ml_usecase='regression',\n", " numerical_features=[], target='charges',\n", " time_features=[])),\n", " ('imputer',\n", " Simple_Imputer(categorical_strategy='not_available',\n", " numeric_strategy='mean',\n", " target_variable=None)),\n", " ('new_levels1',\n", " New_Catagorical_Levels...\n", " ('group', Empty()), ('nonliner', Empty()), ('scaling', Empty()),\n", " ('P_transform', Empty()), ('pt_target', Empty()),\n", " ('binn', Empty()), ('rem_outliers', Empty()),\n", " ('cluster_all', Empty()), ('dummy', Dummify(target='charges')),\n", " ('fix_perfect', Empty()), ('clean_names', Clean_Colum_Names()),\n", " ('feature_select', Empty()), ('fix_multi', Empty()),\n", " ('dfs', Empty()), ('pca', Empty())],\n", " verbose=False)" ] }, "execution_count": 28, "metadata": {}, "output_type": "execute_result" } ], "source": [ "from sklearn import set_config\n", "set_config(display='diagram')\n", "loaded_bestmodel[0]" ] }, { "cell_type": "code", "execution_count": 29, "metadata": {}, "outputs": [], "source": [ "from sklearn import set_config\n", "set_config(display='text')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 14. Deploy Model" ] }, { "cell_type": "code", "execution_count": 30, "metadata": {}, "outputs": [ { "name": "stdout", "output_type": "stream", "text": [ "Model Succesfully Deployed on AWS S3\n" ] } ], "source": [ "deploy_model(best, model_name = 'best-aws', authentication = {'bucket' : 'pycaret-test'})" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 15. Get Config / Set Config" ] }, { "cell_type": "code", "execution_count": 31, "metadata": {}, "outputs": [ { "data": { "text/html": [ "
\n", "\n", "\n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", " \n", "
agebmisex_femalesex_malechildren_0children_1children_2children_3children_4children_5smoker_nosmoker_yesregion_northeastregion_northwestregion_southeastregion_southwest
30036.027.550.01.00.00.00.01.00.00.01.00.01.00.00.00.0
90460.035.101.00.01.00.00.00.00.00.01.00.00.00.00.01.0
67030.031.570.01.00.00.00.01.00.00.01.00.00.00.01.00.0
61749.025.600.01.00.00.01.00.00.00.00.01.00.00.00.01.0
37326.032.900.01.00.00.01.00.00.00.00.01.00.00.00.01.0
\n", "
" ], "text/plain": [ " age bmi sex_female sex_male children_0 children_1 children_2 \\\n", "300 36.0 27.55 0.0 1.0 0.0 0.0 0.0 \n", "904 60.0 35.10 1.0 0.0 1.0 0.0 0.0 \n", "670 30.0 31.57 0.0 1.0 0.0 0.0 0.0 \n", "617 49.0 25.60 0.0 1.0 0.0 0.0 1.0 \n", "373 26.0 32.90 0.0 1.0 0.0 0.0 1.0 \n", "\n", " children_3 children_4 children_5 smoker_no smoker_yes \\\n", "300 1.0 0.0 0.0 1.0 0.0 \n", "904 0.0 0.0 0.0 1.0 0.0 \n", "670 1.0 0.0 0.0 1.0 0.0 \n", "617 0.0 0.0 0.0 0.0 1.0 \n", "373 0.0 0.0 0.0 0.0 1.0 \n", "\n", " region_northeast region_northwest region_southeast region_southwest \n", "300 1.0 0.0 0.0 0.0 \n", "904 0.0 0.0 0.0 1.0 \n", "670 0.0 0.0 1.0 0.0 \n", "617 0.0 0.0 0.0 1.0 \n", "373 0.0 0.0 0.0 1.0 " ] }, "execution_count": 31, "metadata": {}, "output_type": "execute_result" } ], "source": [ "X_train = get_config('X_train')\n", "X_train.head()" ] }, { "cell_type": "code", "execution_count": 32, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "123" ] }, "execution_count": 32, "metadata": {}, "output_type": "execute_result" } ], "source": [ "get_config('seed')" ] }, { "cell_type": "code", "execution_count": 33, "metadata": {}, "outputs": [], "source": [ "from pycaret.regression import set_config\n", "set_config('seed', 999)" ] }, { "cell_type": "code", "execution_count": 34, "metadata": {}, "outputs": [ { "data": { "text/plain": [ "999" ] }, "execution_count": 34, "metadata": {}, "output_type": "execute_result" } ], "source": [ "get_config('seed')" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# 16. MLFlow UI" ] }, { "cell_type": "code", "execution_count": null, "metadata": {}, "outputs": [], "source": [ "!mlflow ui" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "# End\n", "Thank you. For more information / tutorials on PyCaret, please visit https://www.pycaret.org" ] } ], "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": "3.7.4" } }, "nbformat": 4, "nbformat_minor": 2 }