Deploying Time Series Models

Duration: 5 min

This module covers the deployment of time series forecasting models such as ARIMA, SARIMA, Prophet, LSTM, and Transformers. Understanding how to deploy these models is crucial for making real-world predictions and integrating them into production environments.

Deploying ARIMA Models

ARIMA (AutoRegressive Integrated Moving Average) models are widely used for time series forecasting. To deploy an ARIMA model, you need to fit the model on historical data, make predictions, and then integrate these predictions into your application. This involves saving the model parameters and loading them when needed.

import pandas as pd
from pmdarima import auto_arima
import joblib

# Load time series data
data = pd.read_csv('time_series_data.csv', parse_dates=['date'], index_col='date')

# Fit ARIMA model
model = auto_arima(data['value'], seasonal=False, stepwise=True)

# Save the model
joblib.dump(model, 'arima_model.pkl')

# Load the model
loaded_model = joblib.load('arima_model.pkl')

# Make predictions
forecast = loaded_model.predict(n_periods=5)
print(forecast)

Try it in Google Colab:

[123.456, 124.567, 125.678, 126.789, 127.890]

Deploying LSTM Models

LSTM (Long Short-Term Memory) networks are a type of recurrent neural network capable of learning order dependence in sequence prediction problems. To deploy an LSTM model, you need to train the model, save it, and then load it for making predictions. This involves using libraries like TensorFlow or PyTorch.

import numpy as np
import pandas as pd
from tensorflow.keras.models import Sequential
from tensorflow.keras.layers import LSTM, Dense
from tensorflow.keras.models import load_model

# Load time series data
data = pd.read_csv('time_series_data.csv', parse_dates=['date'], index_col='date')

# Prepare data for LSTM
def create_dataset(dataset, look_back=1):
    X, Y = [], []
    for i in range(len(dataset)-look_back-1):
        a = dataset[i:(i+look_back), 0]
        X.append(a)
        Y.append(dataset[i + look_back, 0])
    return np.array(X), np.array(Y)

look_back = 1
dataset = data['value'].values
dataset = dataset.reshape(-1, 1)
train_size = int(len(dataset) * 0.67)
train, test = dataset[0:train_size], dataset[train_size:len(dataset)]
trainX, trainY = create_dataset(train, look_back)
testX, testY = create_dataset(test, look_back)

trainX = np.reshape(trainX, (trainX.shape[0], 1, trainX.shape[1]))
testX = np.reshape(testX, (testX.shape[0], 1, testX.shape[1]))

# Create and fit the LSTM network
model = Sequential()
model.add(LSTM(4, input_shape=(1, look_back)))
model.add(Dense(1))
model.compile(loss='mean_squared_error', optimizer='adam')
model.fit(trainX, trainY, epochs=100, batch_size=1, verbose=2)

# Save the model
model.save('lstm_model.h5')

# Load the model
loaded_model = load_model('lstm_model.h5')

# Make predictions
testPredict = loaded_model.predict(testX)
print(testPredict[:5])

💡 Tip: Ensure that the input data for the LSTM model is properly scaled and reshaped to match the training data format.

Key Concepts

Check Your Understanding