The tl;dr:
Learning how to ride a bike requires first hand experience - and a couple of falls. Books / simulations won't do it. Same idea applies to startups.
Whether you're willing to accept the risks inherent when grabbing the handlebars for the first time + how you react after the first couple of falls will uncover if it is something you want to pursue or run away from.
That said, here are some points. Others may not agree - that's their right.
I love the idea of getting into business. But I have no idea how.
Okay, but at one time you had no idea how to code, correct? You took the initiative, picked up a book, met others who code, and built your talent.
Same thing applies here. You need to read books about startups (read some of the posts here - a lot of good suggestions [try steven blank, eric ries, etc]. Meet with other people doing startups (meetup.com, perhaps?) Do some work to determine whether your idea is truly what your intended market is (again, research will show what methodologies to use). Modify / change as data dictates.
What kind of knowledge do I need to start a business?
Knowledge isn't as important as the ability to learn / interpret the results of your actions. Sure, some intellectual baseline is needed, but everything you believe is correct - idea is great, customers will pay, will only cost me x hours / y dollars - is only a guess. And likely not to survive first contact with your intended customer base.
How do you build up the knowledge base of what it takes to create and run a company?
Universities were built on the notion that studying / deconstructing past successes helps prepare a roadmap for future success. Many today state that startups are an entirely different animal with their own set of conditions. Search out both sides, understand their positions, and form your own opinion. Then you're on your way towards building a knowledge base upon which you can test your assumptions.