The reason that we know that the ABC campaign was effective in Uganda is we have good data on prevalence over time.

But that actually doesn't mean at all that if we gave people more exports, more trade, that that would increase prevalence.

So if I asked you, Why don't you go and find me the prevalence in Burkina Faso in 1991?

But now I want to actually also question that.

But this leaves us with a big gap in our knowledge.

So this is another way to see this.

So if you combine the intuition in this figure with some of the data that I talked about before, it suggests that somewhere between 25 percent and 50 percent of the decline in prevalence in Uganda actually would have happened even without any education campaign.

But it's not just the mortality from these disorders.

And the way that I'm going to do that is,

Depending on your usage scenario, add more access rights as you see fit.

So this is a problem for policy. It was a problem for my research.

There's M data points.

So it turns out, if you use this technique, actually your estimates of prevalence are very close to what we get from testing random samples in the population, but they're very, very different than what UNAlDS tells us the prevalences are.

But because people who died this year of AlDS got it a few years ago, we can use this data on mortality to figure out what HlV prevalence was in the past.

When you divide 4 by 36, you get 0.111, or in percent, 11.1 .

So here is trust from a general social survey of the federal government related to inequality.

We're not just talking about a fluke.

I mean, the whole data was there, you know?

It's not the only place we had any interventions.

Let me quickly go over the proof.

And you can't do a reverse inference here. You can't say that.

It's indigenous to China and the Far East, but given that the prevalence of malaria is here in Africa,

Narrator So to start with some lingo about unsupervised learning.

We subtract 4 x 0.1, which is 0.4, which is 0.6.

But an association The higher people's blood levels of vitamin D are, the less heart disease they have, the less cancer.

If yi 0, then this term falls out, and this one here becomes 1 pi as over here.

So this is a graph of prevalence estimated by UNAlDS, and prevalence based on the mortality data for the years in the late 1990s in nine countries in Africa.

Now, I'm not trying to pick on pancake mix.

So now that we've issued commands to the player, let's try to get some data from it. Useful data to have is what the current time of the player is, how far the user is in the video.

If I type in some junk or forget to validate my passwords, I get an error message.

Easy to find and easy to share

And I think that the answer is, we can look at mortality data, and we can use mortality data to figure out what the prevalence was in the past.

Unfortunately, there's almost no good data on HlV prevalence in the general population in Africa until about 2003.

Gather some meaningful statistics

And institutions can learn from this.

In 2 out of 4 experiments, we saw 1.

Is, it's a conditional probability.

This is a police photo. That's actually my car.

That's the mortality rate of sick people in the hospital.

I now have a more challenging quiz for you in which I give you multivariant data in this case 2 dimensional data.

Note

WHO put together figures using the same diagnostic interviews on random samples of the population to allow us to compare rates of mental illness in each society.

If conversely the outcome is somewhere in this green box and it isn't that unlikely after all.

This shows you in areas with low malaria, medium malaria, high malaria what happens to the number of sexual partners as you increase HlV prevalence.

And here is my answer.