[MUSIC] In this module, we've seen how regression
can be used to predict house prices and also be useful in a wide
range of other applications. So, in the introduction to this course,
Carlos introduced this machine learning pipeline,
where we go from data, which gets shoved into some machine learning method, and
we use that to derive intelligence. Well, let's dig into this block diagram
and expand it, now that we've seen these some of these machine learning
tools, in a little bit more detail. Okay, so now we know what we actually use
to fit our data is some training data set, so that's gonna be our data. And in our housing application, where we're going to predict
the price of some house, the data that we collected was,
we had this table of the house ID, and some set of house attributes,
as well as the house's sales price. And we had this for a whole bunch
of houses in our neighborhood, and we collected this data into some table. So, that represented
our training data set. And then we took that data, and what we
did was shoved it through some feature extractor, which in this case is
a very simple feature extractor, where we just choose some
subset of the house attributes. So, in the examples we looked at,
X, our set of features, represented things, like,
we looked at square feet of the house, and we also looked at number of bathrooms. And we talked about possibly
using more features. Again, we'll talk about that more
in the regression course, but those were two that we
examined in this module. And so then,
what did we do with these features? Well, our goal was to
take these features and have some type of model that led to
a prediction of the house price. Okay, so our output, the intelligence that we're deriving,
is the predicted house price. And we're going to do this for
every house in our training data set. We're gonna take its associated features,
and shove it through this machine learning
model, and predict the house price. And what's the machine learning
model we talked about? Well, in this case, it's regression. That's our specific machine learning
model that we're looking at here. Okay, but remember that this machine
learning model had some set of parameters. Okay, so the parameters we call W. These are the weights on our features. So, for example, it's the weight on square
feet or number of bathrooms, and so on. More technically,
these are called regression coefficients. And we talked about estimating
these parameters from data, so our W hat is our estimate
of these parameters. And how did we talk about doing that? Well, we took our predicted output, so our predicted house price, and
we compared it to the true house price. So, the actual sales price that we
recorded in our training data table. So, Y, here, is our actual sales price for the houses in our training data. And we compare to
the predicted house price. And we use a quality metric to measure
how well we're doing with our prediction, using our model, using W hat as
the parameters of that model. Well, how well are we doing? What was the error metric we talked about? The error metric we talked about was
something called residual sum of squares, where we just sum up the square difference
between the actual house sales price and the predicted house sales price, summing
over all houses in our training data set. Okay, so our quality metric is
gonna take our predictions and our actual house sales observations,
spit out this error, and it's gonna go into a machine learning
algorithm that's gonna be used to update the weights,
update our parameters of our model. And we're gonna talk about this
machine learning algorithm, or different variants of it, a lot more
in the actual course on regression. But this is the overall flowchart for
this machine learning method for our house prediction problem. And this loop here,
where we're taking our predictions, computing our error relative to
the actual house sales prices, and updating the weights or
our model parameters. This process tends to
happen in an iterative way, where we update values again and again. Okay, so if we abstract a way, what we
see is we have some training data set. We have some feature extraction process. We have some machine learning model. And it's gonna produce some intelligence,
which in this case is a prediction. And we're gonna assess the quality of our
intelligence with some quality measure. And we're gonna use that error or
accuracy, depending which way we're thinking about measuring it, to adjust our
model parameters using some algorithm. And we're gonna see this type of flow for
machine learning again and again. In this module,
we've seen how to take our data and derive intelligence using
something called regression, where we have a model that relates
our features to our output. And we talked about this in the context
of predicting house values, and you also work through a really
interesting IPython notebook. And from this, you should be able to
deploy really interesting regression models in practice now. [MUSIC]