msb-public
/
ML-For-Beginners
mirror of https://github.com/microsoft/ML-For-Beginners.git
1
0
Code Issues Projects Releases Wiki Activity
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
Branches Tags
${ item.name }
Create tag ${ searchTerm }
Create branch ${ searchTerm }
from 'f2d23be5f8'
${ noResults }
ML-For-Beginners/7-TimeSeries/2-ARIMA/solution/notebook.ipynb

1089 lines
346 KiB
Raw Normal View History Unescape

time series 1
3 years ago
{
"cells": [
{
editing solution notebook supporting edit for HORIZON loop
3 years ago
"cell_type": "markdown",
time series 1
3 years ago
"source": [
"# Time series forecasting with ARIMA\n",
"\n",
"In this notebook, we demonstrate how to:\n",
time series work
3 years ago
"- prepare time series data for training an ARIMA time series forecasting model\n",
time series 1
3 years ago
"- implement a simple ARIMA model to forecast the next HORIZON steps ahead (time *t+1* through *t+HORIZON*) in the time series\n",
"- evaluate the model \n",
"\n",
"\n",
"The data in this example is taken from the GEFCom2014 forecasting competition<sup>1</sup>. It consists of 3 years of hourly electricity load and temperature values between 2012 and 2014. The task is to forecast future values of electricity load. In this example, we show how to forecast one time step ahead, using historical load data only.\n",
"\n",
"<sup>1</sup>Tao Hong, Pierre Pinson, Shu Fan, Hamidreza Zareipour, Alberto Troccoli and Rob J. Hyndman, \"Probabilistic energy forecasting: Global Energy Forecasting Competition 2014 and beyond\", International Journal of Forecasting, vol.32, no.3, pp 896-913, July-September, 2016."
time series work
3 years ago
],
"metadata": {}
},
inform about required libraries and versions
3 years ago
{
"cell_type": "markdown",
"source": [
"## Install Dependencies\n",
"Get started by installing some of the required dependencies. These libraries with their corresponding versions are known to work for the solution:\n",
"\n",
"* `statsmodels == 0.12.2`\n",
"* `matplotlib == 3.4.2`\n",
"* `scikit-learn == 0.24.2`\n"
],
"metadata": {}
},
time series work
3 years ago
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 16,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"!pip install statsmodels"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
rolling back solution notebook to original example by FL
3 years ago
"/bin/sh: pip: command not found\n"
time series work
3 years ago
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 17,
time series 1
3 years ago
"source": [
"import os\n",
"import warnings\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import pandas as pd\n",
"import datetime as dt\n",
"import math\n",
"\n",
"from pandas.plotting import autocorrelation_plot\n",
"from statsmodels.tsa.statespace.sarimax import SARIMAX\n",
"from sklearn.preprocessing import MinMaxScaler\n",
"from common.utils import load_data, mape\n",
"from IPython.display import Image\n",
"\n",
"%matplotlib inline\n",
"pd.options.display.float_format = '{:,.2f}'.format\n",
"np.set_printoptions(precision=2)\n",
"warnings.filterwarnings(\"ignore\") # specify to ignore warning messages\n"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"outputs": [],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 18,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"energy = load_data('./data')[['load']]\n",
"energy.head(10)"
],
time series work
3 years ago
"outputs": [
{
"output_type": "execute_result",
"data": {
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>load</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2012-01-01 00:00:00</th>\n",
" <td>2,698.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 01:00:00</th>\n",
" <td>2,558.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 02:00:00</th>\n",
" <td>2,444.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 03:00:00</th>\n",
" <td>2,402.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 04:00:00</th>\n",
" <td>2,403.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 05:00:00</th>\n",
" <td>2,453.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 06:00:00</th>\n",
" <td>2,560.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 07:00:00</th>\n",
" <td>2,719.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 08:00:00</th>\n",
" <td>2,916.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2012-01-01 09:00:00</th>\n",
" <td>3,105.00</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
rolling back solution notebook to original example by FL
3 years ago
],
"text/plain": [
" load\n",
"2012-01-01 00:00:00 2,698.00\n",
"2012-01-01 01:00:00 2,558.00\n",
"2012-01-01 02:00:00 2,444.00\n",
"2012-01-01 03:00:00 2,402.00\n",
"2012-01-01 04:00:00 2,403.00\n",
"2012-01-01 05:00:00 2,453.00\n",
"2012-01-01 06:00:00 2,560.00\n",
"2012-01-01 07:00:00 2,719.00\n",
"2012-01-01 08:00:00 2,916.00\n",
"2012-01-01 09:00:00 3,105.00"
editing solution notebook supporting edit for HORIZON loop
3 years ago
]
time series work
3 years ago
},
"metadata": {},
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 18
time series work
3 years ago
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Plot all available load data (January 2012 to Dec 2014)"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 19,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"energy.plot(y='load', subplots=True, figsize=(15, 8), fontsize=12)\n",
"plt.xlabel('timestamp', fontsize=12)\n",
"plt.ylabel('load', fontsize=12)\n",
"plt.show()"
],
time series work
3 years ago
"outputs": [
{
"output_type": "display_data",
"data": {
rolling back solution notebook to original example by FL
3 years ago
"image/png": "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
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/plain": [
"<Figure size 1080x576 with 1 Axes>"
rolling back solution notebook to original example by FL
3 years ago
]
time series work
3 years ago
},
"metadata": {
"needs_background": "light"
}
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
editing solution notebook supporting edit for HORIZON loop
3 years ago
"cell_type": "markdown",
time series 1
3 years ago
"source": [
time series 2
3 years ago
"## Create training and testing data sets\n"
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 20,
time series 1
3 years ago
"source": [
"train_start_dt = '2014-11-01 00:00:00'\n",
"test_start_dt = '2014-12-30 00:00:00' "
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"outputs": [],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 21,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"energy[(energy.index < test_start_dt) & (energy.index >= train_start_dt)][['load']].rename(columns={'load':'train'}) \\\n",
" .join(energy[test_start_dt:][['load']].rename(columns={'load':'test'}), how='outer') \\\n",
" .plot(y=['train', 'test'], figsize=(15, 8), fontsize=12)\n",
"plt.xlabel('timestamp', fontsize=12)\n",
"plt.ylabel('load', fontsize=12)\n",
"plt.show()"
],
time series work
3 years ago
"outputs": [
{
"output_type": "display_data",
"data": {
rolling back solution notebook to original example by FL
3 years ago
"image/png": "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
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/plain": [
"<Figure size 1080x576 with 1 Axes>"
rolling back solution notebook to original example by FL
3 years ago
]
time series work
3 years ago
},
"metadata": {
"needs_background": "light"
}
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 22,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"train = energy.copy()[(energy.index >= train_start_dt) & (energy.index < test_start_dt)][['load']]\n",
"test = energy.copy()[energy.index >= test_start_dt][['load']]\n",
"\n",
"print('Training data shape: ', train.shape)\n",
"print('Test data shape: ', test.shape)"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
editing solution notebook supporting edit for HORIZON loop
3 years ago
"Training data shape: (1416, 1)\n",
"Test data shape: (48, 1)\n"
time series work
3 years ago
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 23,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"scaler = MinMaxScaler()\n",
"train['load'] = scaler.fit_transform(train)\n",
"train.head(10)"
],
time series work
3 years ago
"outputs": [
{
"output_type": "execute_result",
"data": {
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>load</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2014-11-01 00:00:00</th>\n",
" <td>0.10</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 01:00:00</th>\n",
" <td>0.07</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 02:00:00</th>\n",
" <td>0.05</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 03:00:00</th>\n",
" <td>0.04</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 04:00:00</th>\n",
" <td>0.06</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 05:00:00</th>\n",
" <td>0.10</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 06:00:00</th>\n",
" <td>0.19</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 07:00:00</th>\n",
" <td>0.31</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 08:00:00</th>\n",
" <td>0.40</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-11-01 09:00:00</th>\n",
" <td>0.48</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
rolling back solution notebook to original example by FL
3 years ago
],
"text/plain": [
" load\n",
"2014-11-01 00:00:00 0.10\n",
"2014-11-01 01:00:00 0.07\n",
"2014-11-01 02:00:00 0.05\n",
"2014-11-01 03:00:00 0.04\n",
"2014-11-01 04:00:00 0.06\n",
"2014-11-01 05:00:00 0.10\n",
"2014-11-01 06:00:00 0.19\n",
"2014-11-01 07:00:00 0.31\n",
"2014-11-01 08:00:00 0.40\n",
"2014-11-01 09:00:00 0.48"
editing solution notebook supporting edit for HORIZON loop
3 years ago
]
time series work
3 years ago
},
"metadata": {},
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 23
time series work
3 years ago
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Original vs scaled data:"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 24,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"energy[(energy.index >= train_start_dt) & (energy.index < test_start_dt)][['load']].rename(columns={'load':'original load'}).plot.hist(bins=100, fontsize=12)\n",
"train.rename(columns={'load':'scaled load'}).plot.hist(bins=100, fontsize=12)\n",
"plt.show()"
],
time series work
3 years ago
"outputs": [
{
"output_type": "display_data",
"data": {
rolling back solution notebook to original example by FL
3 years ago
"image/png": "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
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
rolling back solution notebook to original example by FL
3 years ago
]
time series work
3 years ago
},
"metadata": {
"needs_background": "light"
}
},
{
"output_type": "display_data",
"data": {
rolling back solution notebook to original example by FL
3 years ago
"image/png": "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
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/plain": [
"<Figure size 432x288 with 1 Axes>"
rolling back solution notebook to original example by FL
3 years ago
]
time series work
3 years ago
},
"metadata": {
"needs_background": "light"
}
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Let's also scale the test data"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 25,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"test['load'] = scaler.transform(test)\n",
"test.head()"
],
time series work
3 years ago
"outputs": [
{
"output_type": "execute_result",
"data": {
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>load</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2014-12-30 00:00:00</th>\n",
" <td>0.33</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 01:00:00</th>\n",
" <td>0.29</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 02:00:00</th>\n",
" <td>0.27</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 03:00:00</th>\n",
" <td>0.27</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 04:00:00</th>\n",
" <td>0.30</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
rolling back solution notebook to original example by FL
3 years ago
],
"text/plain": [
" load\n",
"2014-12-30 00:00:00 0.33\n",
"2014-12-30 01:00:00 0.29\n",
"2014-12-30 02:00:00 0.27\n",
"2014-12-30 03:00:00 0.27\n",
"2014-12-30 04:00:00 0.30"
editing solution notebook supporting edit for HORIZON loop
3 years ago
]
time series work
3 years ago
},
"metadata": {},
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 25
time series work
3 years ago
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"## Implement ARIMA method"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 26,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"# Specify the number of steps to forecast ahead\n",
"HORIZON = 3\n",
"print('Forecasting horizon:', HORIZON, 'hours')"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Forecasting horizon: 3 hours\n"
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 27,
time series 1
3 years ago
"source": [
"order = (4, 1, 0)\n",
"seasonal_order = (1, 1, 0, 24)\n",
"\n",
"model = SARIMAX(endog=train, order=order, seasonal_order=seasonal_order)\n",
"results = model.fit()\n",
"\n",
time series 2
3 years ago
"print(results.summary())\n"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
" SARIMAX Results \n",
"==========================================================================================\n",
"Dep. Variable: load No. Observations: 1416\n",
"Model: SARIMAX(4, 1, 0)x(1, 1, 0, 24) Log Likelihood 3477.239\n",
rolling back solution notebook to original example by FL
3 years ago
"Date: Thu, 30 Sep 2021 AIC -6942.477\n",
"Time: 14:36:28 BIC -6911.050\n",
editing solution notebook supporting edit for HORIZON loop
3 years ago
"Sample: 11-01-2014 HQIC -6930.725\n",
" - 12-29-2014 \n",
"Covariance Type: opg \n",
"==============================================================================\n",
" coef std err z P>|z| [0.025 0.975]\n",
"------------------------------------------------------------------------------\n",
"ar.L1 0.8403 0.016 52.226 0.000 0.809 0.872\n",
"ar.L2 -0.5220 0.034 -15.388 0.000 -0.588 -0.456\n",
"ar.L3 0.1536 0.044 3.470 0.001 0.067 0.240\n",
"ar.L4 -0.0778 0.036 -2.158 0.031 -0.148 -0.007\n",
"ar.S.L24 -0.2327 0.024 -9.718 0.000 -0.280 -0.186\n",
"sigma2 0.0004 8.32e-06 47.358 0.000 0.000 0.000\n",
"===================================================================================\n",
"Ljung-Box (L1) (Q): 0.05 Jarque-Bera (JB): 1464.60\n",
"Prob(Q): 0.83 Prob(JB): 0.00\n",
"Heteroskedasticity (H): 0.84 Skew: 0.14\n",
"Prob(H) (two-sided): 0.07 Kurtosis: 8.02\n",
"===================================================================================\n",
"\n",
"Warnings:\n",
"[1] Covariance matrix calculated using the outer product of gradients (complex-step).\n"
]
}
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"## Evaluate the model"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Create a test data point for each HORIZON step."
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 28,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"test_shifted = test.copy()\n",
"\n",
rolling back solution notebook to original example by FL
3 years ago
"for t in range(1, HORIZON):\n",
editing solution notebook supporting edit for HORIZON loop
3 years ago
" test_shifted['load+'+str(t)] = test_shifted['load'].shift(-t, freq='H')\n",
" \n",
"test_shifted = test_shifted.dropna(how='any')\n",
"test_shifted.head(5)"
],
time series work
3 years ago
"outputs": [
{
"output_type": "execute_result",
"data": {
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>load</th>\n",
" <th>load+1</th>\n",
" <th>load+2</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>2014-12-30 00:00:00</th>\n",
" <td>0.33</td>\n",
" <td>0.29</td>\n",
" <td>0.27</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 01:00:00</th>\n",
" <td>0.29</td>\n",
" <td>0.27</td>\n",
" <td>0.27</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 02:00:00</th>\n",
" <td>0.27</td>\n",
" <td>0.27</td>\n",
" <td>0.30</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 03:00:00</th>\n",
" <td>0.27</td>\n",
" <td>0.30</td>\n",
" <td>0.41</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2014-12-30 04:00:00</th>\n",
" <td>0.30</td>\n",
" <td>0.41</td>\n",
" <td>0.57</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
rolling back solution notebook to original example by FL
3 years ago
],
"text/plain": [
" load load+1 load+2\n",
"2014-12-30 00:00:00 0.33 0.29 0.27\n",
"2014-12-30 01:00:00 0.29 0.27 0.27\n",
"2014-12-30 02:00:00 0.27 0.27 0.30\n",
"2014-12-30 03:00:00 0.27 0.30 0.41\n",
"2014-12-30 04:00:00 0.30 0.41 0.57"
editing solution notebook supporting edit for HORIZON loop
3 years ago
]
time series work
3 years ago
},
"metadata": {},
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 28
time series work
3 years ago
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Make predictions on the test data"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 29,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"%%time\n",
"training_window = 720 # dedicate 30 days (720 hours) for training\n",
"\n",
"train_ts = train['load']\n",
"test_ts = test_shifted\n",
"\n",
"history = [x for x in train_ts]\n",
"history = history[(-training_window):]\n",
"\n",
"predictions = list()\n",
"\n",
"# let's user simpler model for demonstration\n",
"order = (2, 1, 0)\n",
"seasonal_order = (1, 1, 0, 24)\n",
"\n",
"for t in range(test_ts.shape[0]):\n",
" model = SARIMAX(endog=history, order=order, seasonal_order=seasonal_order)\n",
" model_fit = model.fit()\n",
" yhat = model_fit.forecast(steps = HORIZON)\n",
" predictions.append(yhat)\n",
" obs = list(test_ts.iloc[t])\n",
" # move the training window\n",
" history.append(obs[0])\n",
" history.pop(0)\n",
" print(test_ts.index[t])\n",
" print(t+1, ': predicted =', yhat, 'expected =', obs)"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"2014-12-30 00:00:00\n",
"1 : predicted = [0.32 0.29 0.28] expected = [0.32945389435989236, 0.2900626678603402, 0.2739480752014323]\n",
"2014-12-30 01:00:00\n",
"2 : predicted = [0.3 0.29 0.3 ] expected = [0.2900626678603402, 0.2739480752014323, 0.26812891674127126]\n",
"2014-12-30 02:00:00\n",
"3 : predicted = [0.27 0.28 0.32] expected = [0.2739480752014323, 0.26812891674127126, 0.3025962399283795]\n",
"2014-12-30 03:00:00\n",
"4 : predicted = [0.28 0.32 0.42] expected = [0.26812891674127126, 0.3025962399283795, 0.40823634735899716]\n",
"2014-12-30 04:00:00\n",
"5 : predicted = [0.3 0.39 0.54] expected = [0.3025962399283795, 0.40823634735899716, 0.5689346463742166]\n",
"2014-12-30 05:00:00\n",
"6 : predicted = [0.4 0.55 0.66] expected = [0.40823634735899716, 0.5689346463742166, 0.6799462846911368]\n",
"2014-12-30 06:00:00\n",
"7 : predicted = [0.57 0.68 0.75] expected = [0.5689346463742166, 0.6799462846911368, 0.7309758281110115]\n",
"2014-12-30 07:00:00\n",
"8 : predicted = [0.68 0.75 0.8 ] expected = [0.6799462846911368, 0.7309758281110115, 0.7511190689346463]\n",
"2014-12-30 08:00:00\n",
"9 : predicted = [0.75 0.8 0.82] expected = [0.7309758281110115, 0.7511190689346463, 0.7636526410026856]\n",
"2014-12-30 09:00:00\n",
"10 : predicted = [0.77 0.78 0.78] expected = [0.7511190689346463, 0.7636526410026856, 0.7381378692927483]\n",
"2014-12-30 10:00:00\n",
"11 : predicted = [0.76 0.75 0.74] expected = [0.7636526410026856, 0.7381378692927483, 0.7188898836168307]\n",
"2014-12-30 11:00:00\n",
"12 : predicted = [0.77 0.76 0.75] expected = [0.7381378692927483, 0.7188898836168307, 0.7090420769919425]\n",
"2014-12-30 12:00:00\n",
"13 : predicted = [0.7 0.68 0.69] expected = [0.7188898836168307, 0.7090420769919425, 0.7081468218442255]\n",
"2014-12-30 13:00:00\n",
"14 : predicted = [0.72 0.73 0.76] expected = [0.7090420769919425, 0.7081468218442255, 0.7385854968666068]\n",
"2014-12-30 14:00:00\n",
"15 : predicted = [0.71 0.73 0.86] expected = [0.7081468218442255, 0.7385854968666068, 0.8478066248880931]\n",
"2014-12-30 15:00:00\n",
"16 : predicted = [0.73 0.85 0.97] expected = [0.7385854968666068, 0.8478066248880931, 0.9516562220232765]\n",
"2014-12-30 16:00:00\n",
"17 : predicted = [0.87 0.99 0.97] expected = [0.8478066248880931, 0.9516562220232765, 0.934198746642793]\n",
"2014-12-30 17:00:00\n",
"18 : predicted = [0.94 0.92 0.86] expected = [0.9516562220232765, 0.934198746642793, 0.8876454789615038]\n",
"2014-12-30 18:00:00\n",
"19 : predicted = [0.94 0.89 0.82] expected = [0.934198746642793, 0.8876454789615038, 0.8294538943598924]\n",
"2014-12-30 19:00:00\n",
"20 : predicted = [0.88 0.82 0.71] expected = [0.8876454789615038, 0.8294538943598924, 0.7197851387645477]\n",
"2014-12-30 20:00:00\n",
"21 : predicted = [0.83 0.72 0.58] expected = [0.8294538943598924, 0.7197851387645477, 0.5747538048343777]\n",
"2014-12-30 21:00:00\n",
"22 : predicted = [0.72 0.58 0.47] expected = [0.7197851387645477, 0.5747538048343777, 0.4592658907788718]\n",
"2014-12-30 22:00:00\n",
"23 : predicted = [0.58 0.47 0.39] expected = [0.5747538048343777, 0.4592658907788718, 0.3858549686660697]\n",
"2014-12-30 23:00:00\n",
"24 : predicted = [0.46 0.38 0.34] expected = [0.4592658907788718, 0.3858549686660697, 0.34377797672336596]\n",
"2014-12-31 00:00:00\n",
"25 : predicted = [0.38 0.34 0.33] expected = [0.3858549686660697, 0.34377797672336596, 0.32542524619516544]\n",
"2014-12-31 01:00:00\n",
"26 : predicted = [0.36 0.34 0.34] expected = [0.34377797672336596, 0.32542524619516544, 0.33034914950760963]\n",
"2014-12-31 02:00:00\n",
"27 : predicted = [0.32 0.32 0.35] expected = [0.32542524619516544, 0.33034914950760963, 0.3706356311548791]\n",
"2014-12-31 03:00:00\n",
"28 : predicted = [0.32 0.36 0.47] expected = [0.33034914950760963, 0.3706356311548791, 0.470008952551477]\n",
"2014-12-31 04:00:00\n",
"29 : predicted = [0.37 0.48 0.65] expected = [0.3706356311548791, 0.470008952551477, 0.6145926589077886]\n",
"2014-12-31 05:00:00\n",
"30 : predicted = [0.48 0.64 0.75] expected = [0.470008952551477, 0.6145926589077886, 0.7247090420769919]\n",
"2014-12-31 06:00:00\n",
"31 : predicted = [0.63 0.73 0.79] expected = [0.6145926589077886, 0.7247090420769919, 0.786034019695613]\n",
"2014-12-31 07:00:00\n",
"32 : predicted = [0.71 0.76 0.79] expected = [0.7247090420769919, 0.786034019695613, 0.8012533572068039]\n",
"2014-12-31 08:00:00\n",
"33 : predicted = [0.79 0.82 0.83] expected = [0.786034019695613, 0.8012533572068039, 0.7994628469113696]\n",
"2014-12-31 09:00:00\n",
"34 : predicted = [0.82 0.83 0.81] expected = [0.8012533572068039, 0.7994628469113696, 0.780214861235452]\n",
"2014-12-31 10:00:00\n",
"35 : predicted = [0.8 0.78 0.76] expected = [0.7994628469113696, 0.780214861235452, 0.7587287376902416]\n",
"2014-12-31 11:00:00\n",
"36 : predicted = [0.77 0.75 0.74] expected = [0.780214861235452, 0.7587287376902416, 0.7367949865711727]\n",
"2014-12-31 12:00:00\n",
"37 : predicted = [0.77 0.76 0.76] expected = [0.7587287376902416, 0.7367949865711727, 0.7188898836168307]\n",
"2014-12-31 13:00:00\n",
"38 : predicted = [0.75 0.75 0.78] expected = [0.7367949865711727, 0.7188898836168307, 0.7273948075201431]\n",
"2014-12-31 14:00:00\n",
"39 : predicted = [0.73 0.75 0.87] expected = [0.7188898836168307, 0.7273948075201431, 0.8299015219337511]\n",
"2014-12-31 15:00:00\n",
"40 : predicted = [0.74 0.85 0.96] expected = [0.7273948075201431, 0.8299015219337511, 0.909579230080573]\n",
"2014-12-31 16:00:00\n",
"41 : predicted = [0.83 0.94 0.93] expected = [0.8299015219337511, 0.909579230080573, 0.855863921217547]\n",
"2014-12-31 17:00:00\n",
"42 : predicted = [0.94 0.93 0.88] expected = [0.909579230080573, 0.855863921217547, 0.7721575649059982]\n",
"2014-12-31 18:00:00\n",
"43 : predicted = [0.87 0.82 0.77] expected = [0.855863921217547, 0.7721575649059982, 0.7023276633840643]\n",
"2014-12-31 19:00:00\n",
"44 : predicted = [0.79 0.73 0.63] expected = [0.7721575649059982, 0.7023276633840643, 0.6195165622202325]\n",
"2014-12-31 20:00:00\n",
"45 : predicted = [0.7 0.59 0.46] expected = [0.7023276633840643, 0.6195165622202325, 0.5425246195165621]\n",
"2014-12-31 21:00:00\n",
"46 : predicted = [0.6 0.47 0.36] expected = [0.6195165622202325, 0.5425246195165621, 0.4735899731423454]\n",
rolling back solution notebook to original example by FL
3 years ago
"CPU times: user 12min 15s, sys: 2min 39s, total: 14min 54s\n",
"Wall time: 2min 36s\n"
time series work
3 years ago
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {
"scrolled": true
}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Compare predictions to actual load"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 30,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"eval_df = pd.DataFrame(predictions, columns=['t+'+str(t) for t in range(1, HORIZON+1)])\n",
"eval_df['timestamp'] = test.index[0:len(test.index)-HORIZON+1]\n",
"eval_df = pd.melt(eval_df, id_vars='timestamp', value_name='prediction', var_name='h')\n",
"eval_df['actual'] = np.array(np.transpose(test_ts)).ravel()\n",
"eval_df[['prediction', 'actual']] = scaler.inverse_transform(eval_df[['prediction', 'actual']])\n",
"eval_df.head()"
],
time series work
3 years ago
"outputs": [
{
"output_type": "execute_result",
"data": {
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/html": [
"<div>\n",
"<style scoped>\n",
" .dataframe tbody tr th:only-of-type {\n",
" vertical-align: middle;\n",
" }\n",
"\n",
" .dataframe tbody tr th {\n",
" vertical-align: top;\n",
" }\n",
"\n",
" .dataframe thead th {\n",
" text-align: right;\n",
" }\n",
"</style>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th>timestamp</th>\n",
" <th>h</th>\n",
" <th>prediction</th>\n",
" <th>actual</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th>0</th>\n",
" <td>2014-12-30 00:00:00</td>\n",
" <td>t+1</td>\n",
" <td>3,008.74</td>\n",
" <td>3,023.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>1</th>\n",
" <td>2014-12-30 01:00:00</td>\n",
" <td>t+1</td>\n",
" <td>2,955.53</td>\n",
" <td>2,935.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>2</th>\n",
" <td>2014-12-30 02:00:00</td>\n",
" <td>t+1</td>\n",
" <td>2,900.17</td>\n",
" <td>2,899.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>3</th>\n",
" <td>2014-12-30 03:00:00</td>\n",
" <td>t+1</td>\n",
" <td>2,917.69</td>\n",
" <td>2,886.00</td>\n",
" </tr>\n",
" <tr>\n",
" <th>4</th>\n",
" <td>2014-12-30 04:00:00</td>\n",
" <td>t+1</td>\n",
" <td>2,946.99</td>\n",
" <td>2,963.00</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
rolling back solution notebook to original example by FL
3 years ago
],
"text/plain": [
" timestamp h prediction actual\n",
"0 2014-12-30 00:00:00 t+1 3,008.74 3,023.00\n",
"1 2014-12-30 01:00:00 t+1 2,955.53 2,935.00\n",
"2 2014-12-30 02:00:00 t+1 2,900.17 2,899.00\n",
"3 2014-12-30 03:00:00 t+1 2,917.69 2,886.00\n",
"4 2014-12-30 04:00:00 t+1 2,946.99 2,963.00"
editing solution notebook supporting edit for HORIZON loop
3 years ago
]
time series work
3 years ago
},
"metadata": {},
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 30
time series work
3 years ago
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Compute the **mean absolute percentage error (MAPE)** over all predictions\n",
"\n",
"$$MAPE = \\frac{1}{n} \\sum_{t=1}^{n}|\\frac{actual_t - predicted_t}{actual_t}|$$"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 31,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"if(HORIZON > 1):\n",
" eval_df['APE'] = (eval_df['prediction'] - eval_df['actual']).abs() / eval_df['actual']\n",
" print(eval_df.groupby('h')['APE'].mean())"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
editing solution notebook supporting edit for HORIZON loop
3 years ago
"h\n",
"t+1 0.01\n",
"t+2 0.01\n",
"t+3 0.02\n",
"Name: APE, dtype: float64\n"
time series work
3 years ago
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
rolling back solution notebook to original example by FL
3 years ago
{
"cell_type": "markdown",
"source": [],
"metadata": {}
},
time series 1
3 years ago
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 32,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"print('One step forecast MAPE: ', (mape(eval_df[eval_df['h'] == 't+1']['prediction'], eval_df[eval_df['h'] == 't+1']['actual']))*100, '%')"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"One step forecast MAPE: 0.5570581332313952 %\n"
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 33,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [
"print('Multi-step forecast MAPE: ', mape(eval_df['prediction'], eval_df['actual'])*100, '%')"
],
time series work
3 years ago
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Multi-step forecast MAPE: 1.1460048657704118 %\n"
]
}
],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "markdown",
"source": [
"Plot the predictions vs the actuals for the first week of the test set"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
rolling back solution notebook to original example by FL
3 years ago
"execution_count": 34,
time series 1
3 years ago
"source": [
"if(HORIZON == 1):\n",
" ## Plotting single step forecast\n",
" eval_df.plot(x='timestamp', y=['actual', 'prediction'], style=['r', 'b'], figsize=(15, 8))\n",
"\n",
"else:\n",
" ## Plotting multi step forecast\n",
" plot_df = eval_df[(eval_df.h=='t+1')][['timestamp', 'actual']]\n",
" for t in range(1, HORIZON+1):\n",
" plot_df['t+'+str(t)] = eval_df[(eval_df.h=='t+'+str(t))]['prediction'].values\n",
"\n",
" fig = plt.figure(figsize=(15, 8))\n",
" ax = plt.plot(plot_df['timestamp'], plot_df['actual'], color='red', linewidth=4.0)\n",
" ax = fig.add_subplot(111)\n",
" for t in range(1, HORIZON+1):\n",
" x = plot_df['timestamp'][(t-1):]\n",
" y = plot_df['t+'+str(t)][0:len(x)]\n",
" ax.plot(x, y, color='blue', linewidth=4*math.pow(.9,t), alpha=math.pow(0.8,t))\n",
" \n",
" ax.legend(loc='best')\n",
" \n",
"plt.xlabel('timestamp', fontsize=12)\n",
"plt.ylabel('load', fontsize=12)\n",
"plt.show()"
editing solution notebook supporting edit for HORIZON loop
3 years ago
],
"outputs": [
{
"output_type": "stream",
"name": "stderr",
"text": [
"No handles with labels found to put in legend.\n"
]
},
{
"output_type": "display_data",
"data": {
rolling back solution notebook to original example by FL
3 years ago
"image/png": "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
editing solution notebook supporting edit for HORIZON loop
3 years ago
"text/plain": [
"<Figure size 1080x576 with 1 Axes>"
rolling back solution notebook to original example by FL
3 years ago
]
editing solution notebook supporting edit for HORIZON loop
3 years ago
},
"metadata": {
"needs_background": "light"
}
}
],
"metadata": {}
time series 1
3 years ago
},
{
"cell_type": "code",
"execution_count": null,
editing solution notebook supporting edit for HORIZON loop
3 years ago
"source": [],
time series 1
3 years ago
"outputs": [],
editing solution notebook supporting edit for HORIZON loop
3 years ago
"metadata": {}
time series 1
3 years ago
}
],
"metadata": {
"kernel_info": {
"name": "python3"
},
"kernelspec": {
inform about required libraries and versions
3 years ago
"name": "python3",
rolling back solution notebook to original example by FL
3 years ago
"display_name": "Python 3.7.0 64-bit"
time series 1
3 years ago
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
rolling back solution notebook to original example by FL
3 years ago
"version": "3.7.0"
time series 1
3 years ago
},
"nteract": {
"version": "nteract-front-end@1.0.0"
time series work
3 years ago
},
"metadata": {
"interpreter": {
"hash": "70b38d7a306a849643e446cd70466270a13445e5987dfa1344ef2b127438fa4d"
}
inform about required libraries and versions
3 years ago
},
"interpreter": {
rolling back solution notebook to original example by FL
3 years ago
"hash": "70b38d7a306a849643e446cd70466270a13445e5987dfa1344ef2b127438fa4d"
time series 1
3 years ago
}
},
"nbformat": 4,
"nbformat_minor": 2
Changed pip to !pip
3 years ago
}