search

aremazeilles / beat_routine

Performance Indicators developed by BEAT consortium for the Eurobench project
0 stars 2 forks source link

Example of some experimental data for whole system benchmark. #31

Open aremazeilles opened 4 years ago

aremazeilles commented 4 years ago

We will be soon launching an evaluation of our benchmarking website, and as we progressed well with your code, we could eventually use it already for making this evaluation. To do so, we would need some examples of data recorded during one of your protocols, which could be uploaded to the system as if it was an experiment conducted that needs to be scored.

Do you think that would be feasible? It might be better to discuss through a telco, but what we would need is the set of data that would get collected during the execution of one of the protocol, and which would correspond to the data that should be uploaed to the system to permit later on to score the experiment. It could be good to have it for a couple of experiments, so that we could conduct some "experiment comparison".

If you would have already some data collected, we can discuss it.

juritaborri commented 4 years ago

We are planning to conduct the experimental protocols with the Hospital (partner of the project) the first week of the November. If you want, after that we can have a telco also with Fabrizio, the coordinator of the project.

aremazeilles commented 3 years ago

Any update on this? Appart from some reference data, it would be great sharing with us the naming of the files collected data.

We proposed a naming structure in the documentation, we would like to check if all testebds provide pre-processed data using similar file naming (and if not check how it is).

aremazeilles commented 3 years ago

OK At a first stage, sharing the file naming pattern you use would be a good information. We need to check if it is conform, so that we know how to sequentially process your files.

juritaborri commented 3 years ago

We have actually acquired two subjects and tested them with and without exoskeleton. For example as regards the jointAngle data the file name is:

subject_1_cond_1_jointAngle.csv subject_1_cond_2_jointAngle. csv

where 1 and 2 stand for with and without exoskeleton. FOllowing the documentation regarding the file naming, I think it is ok.

aremazeilles commented 3 years ago

Yep it is. The condition with / without exoskeleton could be a variable in the condition files condition_1.yaml and condition_2.yaml , like use_exoskeleton .

aremazeilles commented 3 years ago

In the line with the emails you exchanged, yes it is better, for now, to have one file per condition. In that file, I would rather use boolean flags, so that each of these files could have the following content:

use_exokeleton: yes
is_eyes_open: yes

What do you think?

juritaborri commented 3 years ago

It sound great, ok!

Il Gio 21 Gen 2021, 14:15 Anthony Remazeilles notifications@github.com ha scritto:

In the line with the emails you exchanged, yes it is better, for now, to have one file per condition. In that file, I would rather use boolean flags, so that each of these files could have the following content:

use_exokeleton: yesis_eyes_open: yes

What do you think?

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/aremazeilles/beat_routine/issues/31#issuecomment-764634914, or unsubscribe https://github.com/notifications/unsubscribe-auth/APD2ILILBU4H7LFNNXJHUSTS3ASF3ANCNFSM4SOMZMIQ .