Moving Average

The Moving Average (MA) model is a time series forecasting method that models the output variable as a linear combination of past forecast errors.

Definition

The Moving Average model of order $q$ is denoted as MA(q) and is defined as:

$X_t = \mu + \epsilon_t + \theta_1 \epsilon_{t-1} + \theta_2 \epsilon_{t-2} + \cdots + \theta_q \epsilon_{t-q}$

Where:

• $X_t$ is the actual value at time $t$
• $\mu$ is the mean of the series
• $\epsilon_t$ is the white noise/error term at time $t$
• $\theta_1, \theta_2, ..., \theta_q$ are the model parameters

Key Characteristics

• The MA model depends only on past error terms, not past actual values.
• It is most useful when a time series shows short-term shocks or randomness.
• The autocorrelation function (ACF) of an MA(q) model will cut off after lag $q$.

Example: MA(1)

For a simple Moving Average model of order 1:

$X_t = \mu + \epsilon_t + \theta_1 \epsilon_{t-1}$

This implies the current value depends on the current error and the previous error term.

When to Use

• When residuals (errors) from an AR model still show autocorrelation.
• When the ACF plot shows a sharp cutoff but PACF tails off.

---

Visualization

You can visualize MA processes by plotting synthetic time series data or fitting a model using tools like statsmodels in Python.

from statsmodels.tsa.arima_process import ArmaProcess
import matplotlib.pyplot as plt

# Define an MA(1) process
ma = ArmaProcess(ma=[1, 0.7])  # MA(1) with theta1=0.7
simulated_data = ma.generate_sample(nsample=100)

plt.plot(simulated_data)
plt.title("Simulated MA(1) Process")
plt.xlabel("Time")
plt.ylabel("Value")
plt.grid()
plt.show()