Control the amount of data used in process statistics calculations

In addition to a characteristic’s Filter by options settings, described in the previous topic, the following system parameters can be used to control the amount of sample data that is included in the calculation of process statistics. These parameters are included in the Display group of the General Parameters module.

Cut-off time in days to limit the number of samples while calculating process statistics from the samples (0 = Include all samples)

Upper bound on the age of sample data used to calculate process statistics. This time range is applied to the sample’s requested time. Use this parameter to exclude sample data that is considered too old to be relevant. For example, setting this parameter to a value of 2 would mean that only the last 2 days of sample data would be used to calculate process statistics.

Number of samples to consider calculating process statistics

Maximum number of samples to include when calculating process statistics. Use this parameter to tune the performance of the system so that there is not too much data being analyzed.

These two parameters do not affect the calculation of control limits from the data, which is based on the Samples for Control Limit setting in the characteristic configuration.

These two parameters are used when calculating the statistics that are recorded in the stats table. They can help you to control system performance and ensure that statistic calculations use only recent data. The Cut-off time in days parameter allows you to exclude data that is no longer relevant. The Number of samples to consider parameter allows you to tune the performance of the system by placing a cap on how much data will be included in the process statistics calculation. Logically, they would be used together as follows:

• Perform the process statistics calculations on the top <Number of samples to consider> samples from those that were recorded in the last <Cut-off time in days> days.

For example, if 10 samples per day are being recorded, the cut-off time is set to 2 days, and the number of samples to consider is set to 12, then only the most recent 12 samples will be included in the process statistics calculation. However, if the number of samples to consider is set to 30, and there are 2 days worth of samples recorded, then the most recent 20 samples will be included in the process statistics calculation.