Modify Field Dialog: Correlation Analysis
The Modify Field Dialog allows you to analyze and modify the properties of a field. The Correlation Analysis page allows you to view the results of a correlation analysis produced by the Correlation Analysis Wizard.
Correlation Analysis is useful for searching for direct relationships between values of one data field and previous values of another field. Specifically, correlation analysis determines if the two sets of values move in the same direction by the similar amounts. If a strong relationship is found, the previous values of the second field may be useful for predicting new values of the first.
For example, if there is a high correlation between the closing price of ABC and the closing price of DEF on the previous day, then the changes in the price of DEF may make an effective input for predicting the price of ABC.
Understanding the Significance of the Results
Correlation analysis is a statistical measure of whether two sets of values move in the same direction by the same amount. Like other statistical measures, it is only as good as the data being analyzed. Even if the date range being analyzed has the same characteristics as the data you will be trying to predict, it may still include variability or "noise" that may appear as a correlation when one does not exist.
Because of this, TradingSolutions indicates the values where various levels of confidence exist. For example, the line that indicates "Value needed for 99% confidence" displays the correlation value required to be 99% certain that a correlation exists. This is statistics’ way of saying that 99% of the time, this correlation value was not caused by the values appearing to correlate by chance.
Analyzing Correlation Analysis Results
After the Correlation Analysis Wizard has been run, the results of the correlation analysis are displayed on the Correlation Analysis page. By default, the results are displayed sorted on the absolute value of the correlation. The following information is provided:
p Output Field (or Input Field)
This value indicates the data field for which the correlation analysis was run.
Ä Note: This column is only displayed when correlation analysis is run for a portfolio or subgroup field.
p Input Field (or Output Field)
This value indicates the field for which correlation is being analyzed. This value, combined with the lag size, indicates the value being analyzed.
p Lag
This value indicates the number of samples the Input Field column was lagged before being analyzed. For example, for daily data, a lag of 3 indicates that the correlation values on this line are for the values three trading days before the current value.
p Samples
This value indicates the number of pairs of values that were included in this correlation analysis.
p Correlation
This value indicates the statistical correlation between the data field and the field being analyzed at the given lead/lag length. Specifically, this value is the correlation coefficient between the two sets of data.
The coefficient has a range of 1 to -1:
correlation = 1 perfect correlation
correlation = 0 no correlation found
correlation = -1 inverse correlation
Ä Note: When searching for potential inputs, a high inverse correlation can be just as useful as a high normal correlation. When an input field has a high inverse correlation to the output field, the output field will typically go in the opposite direction as the input field, rather than in the same direction.
p Abs. Value (Absolute Value)
This value displays the absolute value of the correlation. It is used to display the correlation values in order of strength, while disregarding the direction of the correlation.
Ä Note: If more than 200 combinations of data fields, leading fields, and lead/lag values have been analyzed, only the highest 100 and lowest 100 correlation values will be displayed. The lowest 100 correlation values are included since high inverse correlation values can be as useful as high normal correlation values.
Creating Predictions Based on Highly Correlated Values
Using the information in the correlation analysis, highly correlated values can be used to create a prediction of future values of the data field. Depending on the type of correlation analysis that was performed, different options will be available for creating predictions directly from this page.
Create a New Prediction from the Selected Fields…
This button is displayed when correlation analysis is used to determine if other fields may be leading indicators for this data field. To begin creating a prediction, select one or more combinations of leading fields and lead lengths from the display and press this button. This will display the Predict a Value Wizard.
Ä Note: Multiple lines in the display can be selected with the mouse by holding down the Ctrl key when selecting new lines. The Shift key can also be used to select groups of lines.
Create a New Prediction for the Selected Field…
This button is displayed when correlation analysis is used to determine if this data field may be a leading indicator to other fields. To begin creating a prediction, select a combination of lagging field and lag length from the display and press this button. This will display the Predict a Value Wizard.
Display Correlation for Selected Data Field Only…
This button is displayed when correlation analysis is run for an entire group of data fields. Before a prediction can be created from this correlation analysis, the display must be limited to the correlation values for only one data field. To do this, select a data field from the display and press this button. This will display the Modify Field Dialog: Correlation Analysis page (this page) for the selected data field. One of the above buttons will then be displayed for creating predictions.
Ä Note: Any previous correlation analysis data for the selected data field will be replaced when this option is used.
Changing the Correlation Analysis Options
To change the correlation analysis to analyze different fields, a different date range, or a larger maximum lead, press the Change Analysis… button to change the analysis options. This will display the Correlation Analysis Wizard: Analysis Fields and Analysis Options pages. After the options are changed, the correlation analysis will be rerun and the display will be updated to reflect the new analysis.
Deleting Previous Correlation Analysis Results
To delete previous correlation analysis results, press the Delete Analysis button. This is useful for reducing the size of your portfolio files when the results are no longer needed.
What Do I Do Next?
When you are finished updating the properties of the field, press the OK button to save your changes or select another property page. If you would like to exit from this dialog without performing any changes, press the Cancel button.
Additional Property Pages
The Modify Field Dialog has many different property pages that will be displayed as is appropriate for the type of field being modified. For information on the other property pages available, see the help for the Modify Field Dialog.
How Did I Get Here?
The Modify Field Dialog appears when you select to modify or analyze the selected field from the Modify Data Series Dialog or the Modify Portfolio/Group Dialog. It is also displayed when you select Modify Field… from the context menu for a field in a chart or spreadsheet.
The Correlation Analysis page is only displayed when the Correlation Analysis Wizard has been run.