410.20 a

#400#ML_Engineer_Basic#410#Mathematics#410.20#Statistics#410.20 a#Probability_Distribution_and_Estimation

#Statistics
Infer the probability distribution of a parameter from a finite number of data -> It is difficult to know exactly the true probability distribution, and approximate it.

If the structure and number of models change flexibly depending on the data, nonparametric

• A Representative example - Binary data -> Bernoulli distribution
• More than n discrete values -> category distribution
• Data value between 0 and 1 -> Beta distribution
• Data values above 0 -> Gamma distribution, Log normal distribution
• If the data has values throughout R -> Normal distribution Laplace distribution