Knowledge Base

Statistics for Numerical Analysis in Detail

For Numerical Modifiers Only

• Total Duration – The sum of the durations of a scored modifier.
• Total Value – The sum of the values of a scored modifier.
• Mean (Per Minute) – The mean value of a scored modifier weighted for the duration it lasted. For example, modifier 3 lasted 5 seconds and modifier 2 lasted 2 seconds. The Mean (per minute) is ((3 x 5) + (2 x 2)) / (5 + 2) = 19 / 7 = 2.71. The Mean (per minute) is calculated for events with duration only, not for events without duration. If a numerical modifier is associated with an event without duration, the Mean (per minute) is zero.

For Numerical Modifiers and External Data

• Minimum Value – The lowest value of a scored numerical modifier or external data.
• Maximum Value – The highest value of a scored numerical modifier or external data.
• Mean Value – The arithmetic mean of a scored numerical modifier or external data.
• 25th Percentile – Twenty-five percent of the modifiers or external data samples of this type has this value or lower. The 25th percentile is also called the first quartile.
• Median – Fifty percent of the modifiers or external data samples of this type has this value or lower. The Median is the same as the 50th percentile or 2nd quartile.
• 75th Percentile – Seventy-five percent of the modifiers or external data samples of this type has this value or lower. The 75th percentile is also called the 3rd quartile.

For External Data Only

• Number of Samples – The number of external data samples that are included in an interval. This number includes the samples that have a NaN value (missing samples).
• Number of Valid Samples – The number of external data samples that are included in an interval and have a valid value. The number of valid samples is the Number of samples minus the number of samples with a NaN value (missing samples).

Notes

• For external data, all statistics, except for the Number of samples, are based on the Number of valid samples. Hence, missing samples do not influence them.
• Do not group categories if you want to calculate statistics for the external data. Doing so may give unexpected results.

Examples of Calculation of External Data Statistics

The calculation of statistics for external data is based on the actual sample points; the external data is not interpolated.

Example 1: You have an observation with external data. In a data profile, you created intervals based on observation time/event log data. For this interval, the numerical statistics values of the external data are: Minimum = '98', Maximum = '100', Mean = '98.77' (the mean of 98, 100, and 98.3), Number of samples = '3'.

Example 2: You have the same observation with external data as in the previous example. In a data profile, you created an interval by manual selection with external data values higher or equal to 92. For this interval, the numerical statistics values for the external data are: Minimum = '92', Maximum = '100', Mean = '97.07' (the mean of 92, 98, 100, and 98.3), Number of samples = '4'.