Вы находитесь на странице: 1из 24

http://cmapspublic2.ihmc.us/rid=1KZ272GND-1W3JG5F-353V/MDM%20Introduction.

cmap
2 types of random variables (discrete and
continuous), we have 2 types of
distributions:
1. Discrete probability distributions, and
2. Continuous probability distributions
(also known as probability density function,
or density function, or just density).

Examples of discrete probability


distributions :
• Uniform
• Binomial
• Poisson
• Hypergeometric
• Bernoulli
• Geometric

Examples of continuous probability


distributions :
• Uniform
• Normal
• Exponential
• Gamma
• Beta
• Cauchy
• Lognormal
• Double exponential
• t distribution
• F distribution
• Chi-square distribution
Hypothesis Testing provides an approach to making inference about a
population on the basis of sample information. Decisions are made
under uncertainty. The idea behind hypothesis testing is to develop a
statistical tool to minimize the probability of error, to make inferences
about an unknown parameter Mu using the value of the Sample Mean.
The objective of the test is to show evidence against the null. The typical
analogy is when evidence is presented against a defendant, the null is
that the defendant is innocent. It consists of selecting a sample to check
the validity of a conjecture or hypothesis. It provides an approach to
making inference about a population on the basis of sample information.

Elements of a test of hypothesis


• Null Ho
• Alternative H1
• Test statistic
• Rejection region (or level of significance)
• Assumptions about population
• Sample and compute the test statistic
• Perform the test
• Conclusion of the test

A test statistic is the point estimate of the parameter of interest,


obtained from a sample. It is also called the sample statistic. For
instance, if the analyst is interested in the mean of a certain population
the sample statistic is the sample mean, which is referred as the test
statistic. In order to perform a test of hypothesis this test statistic is
standardized. Therefore, the standardized test statistic for the mean of
the population becomes: Standardized test statistic = (Sample mean -
Value from Ho)/ Standard deviation of sample mean

The p-value of a hypothesis test is the smallest value of alpha


(significance level) that would lead to rejection of the null hypothesis. It
is the observed level of significance. A small p-value provides evidence in
support of the alternative hypothesis. In practice, the analyst sets the
desired level of significance to, say, 5%. If the observed significance level
(the p-value) is less than the desired level of significance then the null
hypothesis can be rejected.

Type I error is the probability of overreacting; Reject a true Ho,


Type II error the probability of underreacting; Do not reject a false Ho
Hypothesis Test For the Mean (Normally Distributed Data)
Hypothesis Test For The Median (Non-normal Data)

Вам также может понравиться