search

proksee-project / proksee-cmd

Repo for Proksee Cmd Line Tools
Apache License 2.0
21 stars 2 forks source link

Dev/documentation #43

Closed asahaman closed 3 years ago

asahaman commented 3 years ago

I have added components for machine learning programs in the developer docs of proksee-cmd (this repository). I believe more details on data generation, curation and model building are warranted but it might be fair to have more details in a separate repository (proksee-database in progress). Nevertheless, a shorter version still needs to be correct and clear to the audience and hence also asking @emarinier @jubileepower to review the current write-up.

asahaman commented 3 years ago

"The MachineLearningAssemblyQC class subsequently normalizes the assembly attributes generated from AssemblyQuality class and returns the machine learning probabilistic evaluation of the assembly quality." : This sounds to be for the new isolate/data as input., if so, please note that in the sentence, as the previous paragraph is talking about the model and loading of that.

I agree, how about?

The MachineLearningAssemblyQC class subsequently processes the data of a new/input assembly, by normalizing the assembly attributes generated from AssemblyQuality class and returning the machine learning probabilistic evaluation of the assembly quality.

jubileepower commented 3 years ago

"The MachineLearningAssemblyQC class subsequently normalizes the assembly attributes generated from AssemblyQuality class and returns the machine learning probabilistic evaluation of the assembly quality." : This sounds to be for the new isolate/data as input., if so, please note that in the sentence, as the previous paragraph is talking about the model and loading of that.

I agree, how about?

The MachineLearningAssemblyQC class subsequently processes the data of a new/input assembly, by normalizing the assembly attributes generated from AssemblyQuality class and returning the machine learning probabilistic evaluation of the assembly quality.

Sounds good, thanks!