2.2.2.2 Cross Correlation (CCF)

CCF

Summary

This CCF (Cross Correlation function) tool is supported in the Time Series Analysis App. It is used to compute and plot the cross correlations between two series.

Tutorial

This tutorial uses App’s built-in sample project. To open this sample OPJU file:

1. Right click the Time Series Analysis App icon in the Apps Gallery and choose Show Samples Folder.
2. A folder will open. Drag-and-drop the project file TSA Sample.opju into Origin.

CCF

1. Expand Project Explorer docked on the left. Select folder Statistics and Test . The Book5 contains data about marketing spending and income .
   
2. Highlight Column B and C, and then click the Time Series Analysis App icon in the Apps Gallery.
3. In the dialog, select Statistics and Test and CCF tool.
   
4. Set Number of Lags to 10. Then click the OK button.
   
5. Then you will get the report with CCF graph.
   

Algorithm

Cross correlation calculates the correlation between two time series with different lags. It can be used to identify lags of one time series so that it can be the predictor of the other time series.

This app calls nag_tsa_cross_corr (g13bcc) function [1] to calculate.

Given two time series x_i, y_i, i = 1, 2, ..., n, the cross correlations between x_t and y_t with lags can be calculated as:

where and are means of time series x_i and y_i, s_x and s_y are standard deviations of time series x_i and y_i,

and similarly for s_y.

For lag k<0, it can be calculated by exchange time series x_i and y_i in Equation (1).

• Default Number of Lags

• Standard Error of Cross Correlation

• Confidence Limits

Lower confidence limit at lag k:

Upper confidence limit at lag k:

Reference

1. nag_tsa_cross_corr (g13bcc)