Documente Academic
Documente Profesional
Documente Cultură
Probability
Interpretation of Probabilities:
1. Relative frequency: The chance of having some event can be estimated based on
repeated observation of outcomes. ep: The probability of raining tomorrow, can be
estimated based on number or raining days in historical days with similar climate
conditions.
2. Personal probability: The chance can only be estimated based on one's personal belief.
ep: The chance of giving up, when you failed three times in a puzzle game.
Equally Likely Outcomes: In some experiment, all outcomes have the same relative frequency in the
long run. Such as tossing a fair coin, heads and tails are equally likely to appear. In this case, we can
obtain the probability of each outcome directly. Suppose there are N equally likely outcomes in an
experiment, then each outcome has the probability 1/N to happen.
Events: In an experiment, we do not merely care about the probability of each single outcome. For
example, when tossing a dice, we are not only interested in “something” like whether the outcome
is 6, but also probabilities of other “thing”, such as whether the outcome is an even number. To
abstract this idea, we define probabilities for “events”, which can be thought of as combinations of
outcomes.
Conditional Probabilities
Independence revisited: If A and B are independent events, then P(A&B) = P(A)P(B). By (**), we have
P(A|B) = P(A), which is a more intuitive description of independence. It says, the occurrence of B
does not affect the chance of having A.
Hypothesis test
Choice of Hypothesis: One should choose the hypothesis under which all probabilities can be
calculated.
P-value: The probability of having an outcome that is more extreme or as extreme as the present
one. It is usually calculated using the addition rule, as we can easily see the two cases are mutually
exclusive.
Critical Value: A small P-value means the present outcome is a strong evidence against the
hypothesis. (You might want to spend a few minutes here to think of the logic. In philosophical or
mathematical logic terms, this is called “proof by contradiction”) However, how small is enough? To
be able to make judgement, we choose critical values to compare with P-value. When P is less than
the critical value, we say that the hypothesis is rejected. Common critical values are 0.10, 0.05, 0.01.