Data mining-statistical analysis (5: Statistics)

Statistics

By extracting samples from the population to construct appropriate statistics, the sample function of the population is inferred from the sample nature.

Common statistics

1) average sample value

2) sample variance and standard deviation

3) coefficient of variation of samples-c = root code (d (x)/E (X)

4) K Degree

5) center distance

6) Sample skewness

7) Sample Kurtosis

Note: skewness and kurtosis are widely used in quality control and reliability research.

Order Statistics

Important Statistics of parameter estimation and hypothesis test.

1) Minimum and Maximum Order Statistics

2) sample range = maximum statistic-Minimum statistic

3) median, quantile, and quartile

Full statistics

A statistic that does not lose any information during statistic processing is called a full statistic.

Several Concepts of distribution sampling distribution

If the overall distribution type is known, a mathematical expression of the distribution of a statistic can be derived for any natural number n. The exact distribution is mostly obtained in normal conditions.

Progressive Distribution

When N of the sample distribution is large, the extreme distribution is used as an approximation of the sample distribution. This extreme distribution is called a progressive distribution.

Approximate distribution obtained by Random Simulation

Distribution of observed values satisfied by repeated experiments.

Several Important Distributions derived from normal distribution, Chi-square distribution, tdistribution, and F distribution

It plays an important role in variance analysis and regression equation significance test.

 

 

Data mining-statistical analysis (5: Statistics)