sensing machine diagnostic data, including vibration data, using one or more sensors of a monitoring unit affixed to the machine;

deriving at said monitoring unit a prescribed set of indicators from the machine diagnostic data;

for a first set of said machine diagnostic data, generating at said monitoring unit a fault severity probability for each one of a first plurality of said diagnostic conditions, wherein said generating comprises implementing a resident Bayesian probability model configured with a set of parameters defined for a prescribed set of symptoms for said first plurality of said diagnostic conditions, each one symptom of the prescribed set of symptoms comprising at least one indicator of said prescribed set of indicators;

after said generating, uploading the prescribed set of indicators and the generated fault severity probability for each one of the first plurality of said diagnostic conditions to a machine diagnostic database remote from said machine;

displaying at a first computing device having access to the machine diagnostic database a diagnostic health matrix, including the uploaded fault severity probability generated for a first condition among the first plurality of diagnostic conditions;

receiving as an input at the first computing device an edit to the diagnostic health matrix;

storing at the machine diagnostic database said edit to achieve a corrected fault severity probability for the first condition;

modifying a copy of the Bayesian probability model so that the modified Bayesian probability model provides, in response to said first set of machine diagnostic data, a modified fault severity probability for said first condition that is equal to said corrected fault severity probability, wherein said modifying comprises said learning;

downloading the modified copy of the Bayesian probability model to said monitoring unit to update the resident Bayesian probability model;

repeating said steps of sensing and deriving to derive a newly-derived set of indicators from newly-collected machine diagnostic data;

based on the newly-derived set of indicators and the newly-collected machine diagnostic data, regenerating at said monitoring unit the fault severity probability for each one of the first plurality of said diagnostic conditions using the updated resident Bayesian probability model;

uploading to the machine diagnostic database the regenerated fault severity probability for said first condition; and

displaying, at a second computing device having access to the machine diagnostic database, a warning associated with the first condition for which the corresponding regenerated fault severity probability is suggestive of the maintenance need for the machine.