Consider the following dataset including characteristics of houses and whether or not an agent was able to sell the house:   house color size      sold red large     yes red large     no gray small no gray large     yes gray small yes brown small no brown large     yes   If we were developing a naive bayes model to predict house sales, what is the probability a house would be red given that it was sold?

Which machine learning technique does not produce a model that is a generalization of its input data (instead retains the individual inputs)?

What can we say about ‘nature’, from an agent perspective?

The explore/exploit problem is unqique to ______ learning.

If a machine learning model were predicting coronavirus infections and correctly diagnosed 50 patients with coronavirus, incorrectly diagnosed 22 patients with coronavirus who were healthy, correctly declined to diagnose 500 healthy patients with coronavirus, and missed diagnosing 200 patients with coronavirus when those patients were indeed ill with the disease, what is the number of false-positives by this model in this testing scenario?

Which of the following search strategies are considered ‘uninformed’?

Given this knowledge base:c ← e.d ← k.e.f ← j ∧ e.j ← c.   d can be proven.  

The ratio of true-positive predictions to total positive predictions is known as ____.

If it is known that a goal is only a few arcs away from the starting position in a very large search space with a low branching factor, which search strategy would run the highest risk of having a long or indefinite runtime?

If KB is a set of clauses and g is a conjunction of atoms, g is a ____ of KB, written KB |= g , if g is true in every model of KB.