3 ms·

Here is a much simpler approach that works for many problems: minimize the distance to a desired multi-variate state. https://medium.com/@justchap/using-the-p

by orasis 10y ago

Here is a much simpler approach that works for many problems: minimize the distance to a desired multi-variate state. https://medium.com/@justchap/using-the-pythagorean-theorem-to-model-complicated-goals-in-machine-learning-b85f04b34ad4#.9tajh1t0l https://medium.com/@justchap/using-the-pythagorean-theorem-t...

mccourt 10y ago
That strategy can be viable; it's discussed in the Wikipedia article: https://en.wikipedia.org/wiki/Multi-objective_optimization#No-preference_methods https://en.wikipedia.org/wiki/Multi-objective_optimization#N... As is suggested there, though, implementing this no-preference strategy requires some clean rescaling of the component functions in order to yield equal significance for all of them. If you have such a rescaling, that's outstanding; however, as I suggested in the section of the article dealing with the impact of the choice of currency, rescaling a problem may be a difficult proposition. This is especially true for problems that aren't as simple as the toy problem I've proposed here.