1) HeLa cells are cancer cells. As such, they do not behave like normal, non-cancerous human cells. Some of their physiology is similar, but enough is different that they can't be relied upon for everything.
2) Normal human cells are not always easy to obtain in numbers sufficient to do experiments, especially primary-isolated human cells. There are regulatory issues, ethical issues, and commercial suppliers charge an arm and a leg for them. Finally, normal human cells often change their phenotype once they're in culture, and those changes can be magnified depending on how many "passages" they've gone (and commercially available human cells are often sold at 4th or 5th passage, sometimes even higher, becaused that's the only way to get large numbers of them). Yeast don't generally have that kind of problem. There's other issues, but that one is not the problem that it is with human cells.
3) Yeast biochemistry and physiology is similar enough to that of human cells that they can be very useful. Moreover, growing yeast is cheap and easy. Of course, nobody who works with yeast is going to claim that it's a perfect model for human cells--they're not that naive or stupid. But so many cellular factors discovered in yeast have turned out to have mammalian (including human) counterparts, that yeast is viewed as a very useful model for a lot of cellular functions. It's an experimental model, nothing more. Findings in yeast must be viewed as being merely a first step. Eventually, findings must be confirmed using human cells--but the basic strategy is to do a lot of screening work using the cheap, easy model, then that gives you a clearer roadmap into how to examine the issues using the more expensive and difficult human cells.
That's how research operates--it's the art of the doable--so start out simple, then refine things as you progress.