|
<< Click to Display Table of Contents >> Navigation: EXPORT Behavioral Data > Cluster data export > Suppress wrongly given Codes |
It is also possible to use this function to suppress code combinations that have been logged accidentally.
For instance, if mutual exclusive Codes occur in Events that do overlap in time. This can happen, if you did not use the mutual excluding codes option, or if you code your video in several passes, or if behavior cannot be recognized frame accurate.
For example: If the number of accidentally overlapping frames does not exceed 10 frames, the cluster size can be set to “11”. Now all behavioral patterns that do not last for at least 11 frames are ignored and therefore those Code combinations (none existent in real life) are not counted.
Note: Of course, all other Codes lasting less than 11 frames are ignored as well! Therefore, the usability of this function depends on the minimum length of your 'real' Events. This minimum length is calculated in the time DataSet based statistics.
Example
Suppose Code 'A' and 'B' should exclude each other, but two Events with those Codes overlap each other for 2 Frames. The export results, using a 1 frame cluster, results in data like this:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Error |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
A |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
|
|
|
|
|
|
|
|
|
|
|
|
|
B |
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
1 |
|
|
As long as the Cluster size is set to 1, you do get the following cluster count:
Count |
A |
B |
13 |
1 |
0 |
2 |
1 |
1 |
11 |
0 |
1 |
If you set the cluster size to 3 frames, the results looks like this:
Count |
A |
B |
13 |
1 |
0 |
11 |
0 |
1 |
The false, overlapping, part has now been removed, because the over lapping lasted for only 2 frames and the minimal cluster size was set to 3.