As I see it, and, correct me if I'm wrong, but there is a way to do it.
The problem with using pure Uranium or any other readily fissionable element is that, as one element decays, that releases 2 neutrons which can speed up the decay of nearby elements. If you had a ball of U235 or U238, say, the size of a planet or even a small moon, you'd have a cascade reaction and something resembling a bomb, like a small nova, not a star.
There was a question very similar to this, I remember, but I can't find it. I answered it, perhaps badly, but having answered it, I've given it some thought since then.
The way to do it would be a mix of elements. I think Uranium and Tungsten would be best - with more Tungsten than Uranium, I'd guess at least 10 parts to 1, maybe more and preferably Tungsten 182, cause 182, 183, 184 are all stable, so it could absorb neutrons without becoming radioactive itself.
The density of Uranium and Tungsten are quite similar, with Uranium slightly more dense. We know from the temperature of the earth that radioactive decay heats the interior of the planet, so if we have a planet or star sized object that's maybe 3% or 5% uranium, that's going to generate a lot of heat, and likely create circulation and over time, gradually, the Uranium would fall towards the core and as it grows more concentrated, which would happen slowly, you would see an increased rate of fission.
I see no reason why a planet sized ball of Tungsten 182 mixed with Uranium couldn't be star like, though it wouldn't burn as long, but you might be able to create a surface temperature rivaling a star for a million or a few million years or so.
That would never happen in reality, it's just a theoretical possibility.