2024-04-08 07:41 AM - edited 2024-04-08 07:41 AM
Hello,
I'm trying to figure out how Unico calculates it feature values during .arff creation.
Unfortunately, if I calculate the features as described in AN5804, the feature values differ from the Unico output.
I have attached two files, accelerometer samples with five values per axis and the .arff output of Unico for a window size of five (feature mean X/Y/Z/V).
Can someone explain how the feature values are calculated by Unico during .arff creation?
Best regards
Solved! Go to Solution.
2024-04-09 06:52 AM
Hi @He_Ka ,
Welcome to ST Community!
Values are represented in half-precision floating point, how the values of features and filters are calculated in Unico is confidential information.
Note that it is important to use Unico / MEMS Studio to generate the ARFF file because features/filters are calculated exactly as done at runtime by the embedded logic of Machine Learning Core. If you do not want to use GUI-based tools, such as Unico and MEMS Studio, you can use the Python script available on GitHub.
2024-04-09 06:52 AM
Hi @He_Ka ,
Welcome to ST Community!
Values are represented in half-precision floating point, how the values of features and filters are calculated in Unico is confidential information.
Note that it is important to use Unico / MEMS Studio to generate the ARFF file because features/filters are calculated exactly as done at runtime by the embedded logic of Machine Learning Core. If you do not want to use GUI-based tools, such as Unico and MEMS Studio, you can use the Python script available on GitHub.
2024-04-10 04:34 AM
Thank you for your explanation.
The python scripts you mentioned refer to an mlc_configuration_tool in the setup section of the readme. Unfortunately I am not able to locate a file with this name in my Linux based Unico installation. Can you help me with that?
Best regards