Advanced Options Dialog Box
Used to set advanced options for performing the change-point analysis. You should not change these settings unless you understand their use. Most users of Change-Point Analyzer need never change these settings. The values specified will affect all future analysis. They are stored in the system registry and will remain in effect the next time the program is run. The available options are:
Confidence Level For Identifying Candidate Change: The specified confidence level is used for identifying candidate changes. Changes whose confidence levels are greater than or equal to the specified value are added to the initial pool of changes. However, many of these changes will be later eliminated based on the second parameter described below. It must be entered as a percentage (without a '%') and be at least 50 and less than or equal to the confidence level for inclusion in table of changes below. The default is 50. Setting this value low results in a larger pool of candidate points.
Confidence Level For Inclusion in Table of Changes: The specified confidence is used for deciding whether a change is significant . Only changes whose confidence level are greater than or equal to this value are included in the table of changes. It must be entered as a percentage (without a '%') and be at least 50 and less than 100. The default is 90. Common alternatives are 80, 95 and 99.
Confidence for Confidence Interval: The specified confidence is used for determining the confidence intervals for the time of a change. It must be entered as a percentage (without a '%') and be at least 50 and less than 100. The default is 95. Common alternatives are 90 and 99.
Number of Bootstrap Samples: Bootstrapping is used to determine the confidence level of each change and a confidence interval for the time of each change. The accuracy of these two items is determined by the number of bootstrap samples used. The default value is 1,000. This is the minimum number of bootstrap samples that should be used. This number can be increased to improve the confidence level and confidence interval. However, this will also increase the amount of time required to complete the analysis. For small sets of data it is recommended this value increase to 10,000.
Sampling Method: When generating the bootstrap samples, the samples can be generated by sampling the original data either with or without replacement. Bootstrapping is generally performed with replacement. However, in the case of bootstrapping CUSUM charts, sampling without replacement can also be used. Sampling without replacement generally reduces the variation in the bootstraps increasing the significance levels. It is the default method. Because sampling without replacement generates a smaller set of possible outcomes, it should not be used when the number of outcomes is already limited as with small sets of attribute data.
Method of Estimation: There are two approaches to estimating the time of the change. The Mean Square Error (MSE) is the default approach. It is more susceptible to being biased due to other undetected changes. The second is the CUSUM estimator. It is more susceptible to unusual points at the point of change..
Once the options have been specified, click the OK button or press the Enter key to save the results and close the dialog box. Clicking the Cancel button or pressing the Esc key instead restores the analysis options to their former values and closes the dialog box. To set all analysis options back to the default values, click the Default button.
This dialog box is displayed by selecting the Advanced Options menu item, by clicking on the Advanced Options button on the toolbar, or by clicking the Advanced button on the Custom Analysis dialog box.