-
Naïve Bayes classifiers assume:
a. a conditionally discriminative distribution.
b. independence between attributes.
c. conditional independence between attributes.
d. a strong prior.
Correct answer: c
Analysis:
Naïve Bayes classifiers assume that all the attributes are conditionally independent given the classification variable. This drastically reduces the number of conditional probabilities that need be estimated from the data. As a result Naïve Bayes classifiers are less prone to overfitting then other classifiers with small datasets. So, the right answer is conditional independence between attributes. -
Let us assume that we roll two dice, with possible outcomes val(X)={1,2,3,4,5,6} and val(Y)={1,2,3,4,5,6}. We are interested in the highest dice roll, which we denote Z=max(X,Y). Which one of the following is NOT true:
a. P(X|Y)=P(X).
b. P(Y,X)=P(Y) P(X).
c. P(X,Y | Z)=P(X) P(Y) .
d. P(Y|X)=P(Y).
Correct answer: c
Analysis:
We could see that there are two dices and they are the independent exapmles. So, the probability of each possible exapmle is independent.
-
Bayesian Belief Networks represent knowledge as:
a. a set of conditionally independent variables.
b. directed acyclic graphs and conditional probability tables .
c. cyclic graphs with priors.
d. a set of conventionally independent variables.
Correct answer: b
Analysis:
A Bayesian Belief Network is a probabilistic graphical model that represents the knowledge via a directed acyclic graph(the concept of Bayes network). Each node denotes a variable and each arc indicates a dependency between the nodes. A table of conditional probabilities is associated with each node. -
Assume val{H}={h1,h2,h3} mutually exclusive hypothesis. We observe a piece of data D. Which one of the following identities is always true?
a. P(h1|D)=1−P(D|h2)+P(D|h3).
b. P(D|h1)+P(D|h2)+P(D|h3)=1.
c. All other choices are incorrect .
d. P(h1|D)+P(h2|D)+P(h3|D)=1.
Correct answer: b
Analysis:
-
A rare disease has an incidence rate 0.0001. A very accurate test for the disease exists: it has 0.99 sensitivity (i.e. the probability of coming up positive if having the disease) and a 0.999 specificity (i.e. the probability of coming up negative, while not having the disease). Sensitivity and specificity are also called true positive rate and true negative rate, respectively. A person takes the test for the disease and tests positive, what is the actual probability of having the disease?
Correct answer: 0.09009009009009
Analysis:
-
A friend of yours tosses two coins behind your back and tells you that at least one of them is a head (so both could be heads or either one could be head). He also tells you that the first coin is biased with the probability of coming heads being ph1=0.4, while the second one is fair. Is it worth betting a beer that other coin has landed as a head? Compute the probability that both are heads given the information that at least one landed as a head.
Correct answer: 0.5
Analysis:
Machine Learning(3):Quiz3
最新推荐文章于 2019-11-26 04:28:51 发布