Given a new passenger and knowing his or her personal information, we want to predict whether he or she will survive. The options that we explored until now are based on dividing the passengers into groups and identifying the survival rate for each group. For some combinations of features, such as first class female children, we don't have enough data, so we have to use the survival rate of a larger group such as females not belonging to the third class. We are ignoring some details, for instance, the fact that they are children, and in this way we are losing information. Is there a way to estimate the survival rate for any combination of features, regardless of how many passengers we have?

There are many machine learning algorithms that take account of all the features at the same time. In this chapter, we see a very popular algorithm that is the random forest algorithm. It's not the best option in this context, as it performs better when there are much more features...