Voice recognition with mycroft

I’ve less than perfect diction. Tried voice recognition in the 1990’s, tts and stt, with dismal results so have avoided it.

I’d like to teach mycroft my speech patterns so there are fewer mistakes.

I have an RPI3 that’s presently tied up. I have two RPIs and a number of 32 bit laptops running debian wheezy and jessie.

Can I teach mycroft to recognize my words on an RPI with a Samsung Level bluetooth deivce for communicaiton such that what it learns can be transferred to other machines? Eventually hoping to go with an RPI3 and piPLUSE.

Is there a teaching program, e.g. speaking lists that mycroft learns, to facilitate voice recognition? Which elements of the mycroft package would I need?