Occams razor (the idea that the simplest answer is *most likely the best - thus 'parsimony' or economy) is not provable. It is an elegant idea. It tends to hold up to scrutiny as well. But it remains an aesthetic in that all approaches must - to some extent.
Here's my reasoning:
Solipsism can be described in this way: all 'existence' could be the result of tenaciously consistent, constantly recurring figments of a single entity's imagination. Thus, we may all be figments of some greater consciousnesses fantasy, or everything may be figments of my own delusion. As ridiculous as this sounds, this possibility cannot be entirely ruled out.
Therefore, as likely as it seems that empiricism and parsimony render the best (most true) explanations for reality and existence, they remain an aesthetic (method of understanding or world-view chosen in preference to all others).
Also, considering that even empirically based cosmology admits that observations are limited by the 'horizon problem' (the fact that distant information can be billions of years old or never able to reach us at all due to the upper limit of the speed of light), empiricism can never render a comprehensively accurate picture of the universe - even if any human were able to retain all available data and correctly process all significant correlations between said data.
In a few hundred billion years, for example, cosmologists using the same empirical techniques we use now, would be forced to conclude that the universe consists of a single, steady state galaxy. All other evidence would be too far away to reach us at the speed of light - since the universe appears to be expanding faster than that.
Don't get me wrong. I am an empiricist. However, like Einstein, I hold imagination is higher esteem than knowledge. I am not only willing to accept the unknown - I am made more passionately in love with the universe because of it.
*correction - see Jaume's reply below