Re: [Nmlrg] Machine Learning in network - solicitation for use cases

[Jérôme François <[e-mail]>]

Hi Bing,

Le 18/09/2015 05:48, Liubing (Leo) a écrit :
>
> [Bing] Indeed. The trick/art is in feature selection.
> However, feature selection is basically made by man who understands both the application and the machine learning well. We were always wondering, is there any possibility that machine can select features by itself dynamically according to some general/universal methods.
Automated selection of features could work on very generic features I
guess like byte values, bytelength... but avoiding the expert with
knowledge to design targeted features is still very challenging. But I
think most of the time, there are some desciption of the data format we
are relying which is already a good basis to decompose in
attributes/features. Then, feature selection algortihms can be applied.
>
jerome