White noise is an important concept in time series forecasting.

If a time series is white noise, it is a sequence of random numbers and cannot be predicted. If the series of forecast errors are not white noise, it suggests improvements could be made to the predictive model.

In this tutorial, you will discover white noise time series with Python.

After completing this tutorial, you will know:

- The definition of a white noise time series and why it matters.
- How to check if your time series is white noise.
- Statistics and diagnostic plots to identify white noise in Python.

Let’s get started.

## What is a White Noise Time Series?

A time series may be white noise.

A time series is white noise if the variables are independent and identically distributed with a mean of zero.

This means that all variables have the same variance (*sigma^2*) and each value has a zero correlation with all other values in the series.

If the variables in the series are drawn from a Gaussian distribution, the series is called Gaussian white noise.

## Why Does it Matter?

White noise is an important concept in time series analysis and forecasting.

It is important for two main reasons:

**Predictability**: If your time series is white noise, then, by definition, it is random. You cannot reasonably model it and make predictions.**Model Diagnostics**: The series of errors from a time series forecast model should ideally be white noise.

Model Diagnostics is an important area of time series forecasting.

Time series data are expected to contain some white noise component on top of the signal generated by the underlying process.

For example:

1 |
y(t) = signal(t) + noise(t) |

Once predictions have been made by a time series forecast model, they can be collected and analyzed. The series of forecast errors should ideally be white noise.

When forecast errors are white noise, it means that all of the signal information in the time series has been harnessed by the model in order to make predictions. All that is left is the random fluctuations that cannot be modeled.

A sign that model predictions are not white noise is an indication that further improvements to the forecast model may be possible.

### Stop learning Time Series Forecasting the *slow way*!

Take my free 7-day email course and discover data prep, modeling and more (with sample code).

Click to sign-up and also get a free PDF Ebook version of the course.

## Is your Time Series White Noise?

Your time series is not white noise if any of the following conditions are true:

- Does your series have a zero mean?
- Does the variance change over time?
- Do values correlate with lag values?

Some tools that you can use to check if your time series is white noise are:

**Create a line plot**. Check for gross features like a changing mean, variance, or obvious relationship between lagged variables.**Calculate summary statistics**. Check the mean and variance of the whole series against the mean and variance of meaningful contiguous blocks of values in the series (e.g. days, months, or years).**Create an autocorrelation plot**. Check for gross correlation between lagged variables.

## Example of White Noise Time Series

In this section, we will create a Gaussian white noise series in Python and perform some checks.

It is helpful to create and review a white noise time series in practice. It will provide the frame of reference and example plots and statistical tests to use and compare on your own time series projects to check if they are white noise.

Firstly, we can create a list of 1,000 random Gaussian variables using the *gauss()* function from the random module.

We will draw variables from a Gaussian distribution with a mean (*mu*) of 0.0 and a standard deviation (*sigma*) of 1.0.

Once created, we can wrap the list in a Pandas Series for convenience.

1 2 3 4 5 6 7 8 9 |
from random import gauss from random import seed from pandas import Series from pandas.tools.plotting import autocorrelation_plot # seed random number generator seed(1) # create white noise series series = [gauss(0.0, 1.0) for i in range(1000)] series = Series(series) |

Next, we can calculate and print some summary statistics, including the mean and standard deviation of the series.

1 2 |
# summary stats print(series.describe()) |

Given that we defined the mean and standard deviation when drawing the random numbers, there should be no surprises.

1 2 3 4 5 6 7 8 |
count 1000.000000 mean -0.013222 std 1.003685 min -2.961214 25% -0.684192 50% -0.010934 75% 0.703915 max 2.737260 |

We can see that the mean is nearly 0.0 and the standard deviation is nearly 1.0. Some variance is expected given the small size of the sample.

If we had more data, it might be more interesting to split the series in half and calculate and compare the summary statistics for each half. We would expect to see a similar mean and standard deviation for each sub-series.

Now we can create some plots, starting with a line plot of the series.

1 2 3 |
# line plot series.plot() pyplot.show() |

We can see that it does appear that the series is random.

We can also create a histogram and confirm the distribution is Gaussian.

1 2 3 |
# histogram plot series.hist() pyplot.show() |

Indeed, the histogram shows the tell-tale bell-curve shape.

Finally, we can create a correlogram and check for any autocorrelation with lag variables.

1 2 3 |
# autocorrelation autocorrelation_plot(series) pyplot.show() |

The correlogram does not show any obvious autocorrelation pattern.

There are some spikes above the 95% and 99% confidence level, but these are a statistical fluke.

For completeness, the complete code listing is provided below.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 |
from random import gauss from random import seed from pandas import Series from pandas.tools.plotting import autocorrelation_plot from matplotlib import pyplot # seed random number generator seed(1) # create white noise series series = [gauss(0.0, 1.0) for i in range(1000)] series = Series(series) # summary stats print(series.describe()) # line plot series.plot() pyplot.show() # histogram plot series.hist() pyplot.show() # autocorrelation autocorrelation_plot(series) pyplot.show() |

## Further Reading

This section lists some resources for further reading on white noise and white noise time series.

- White Noise Process, Page 28, Time Series Analysis: Forecasting and Control
- Section 4.2 White Noise, Introductory Time Series with R.
- White Noise on Wikipedia
- Gaussian noise on Wikipedia

## Summary

In this tutorial, you discovered white noise time series in Python.

Specifically, you learned:

- White noise time series is defined by a zero mean, constant variance, and zero correlation.
- If your time series is white noise, it cannot be predicted, and if your forecast residuals are not white noise, you may be able to improve your model.
- The statistics and diagnostic plots you can use on your time series to check if it is white noise.

Do you have any questions about this tutorial? Ask your questions in the comments below and I will do my best to answer.

Do I need to remove the trend of data before we check for white noise ? Thanks

Yes, great question!