Test of the normality of the R language

Test of the normality of the R language

Normal distribution Test with R language (2012-02-29-10:59:54) reprint

Excerpt from: Wu Xizhi: Non-parametric statistics (second edition), China Statistical Publishing house, October 2006: p164-165

1, Ks.test ()
For example 0 is assumed to be N (15,0.2), then Ks.test (x, "Pnorm", 15,0.2). If not the normal distribution, you can also choose "Pexp", "Pgamma" and so on.
2, Shapiro.test ()
Shapiro-wilk tests can be conducted on the normal distribution.
3. Nortest Bag
Lillie.test () can carry out a more accurate Kolmogorov-smirnov test.
Ad.test () is tested for anderson-darling normality.
Cvm.test () Cramer-von mises normality test.
Pearson.test () Pearson card for the test of the Fangzheng state.
Sf.test () is tested for shapiro-francia normality.
4. Fbasics Bag
Normaltest () is tested for Kolmogorov-smirnov normality.
Ksnormtest () is tested for Kolmogorov-smirnov normality.
Shapirotest () carries out the Shapiro-wilk ' s normal state test.
Jarqueberatest () is tested for jarque-bera normality.
Dagotest the d ' Agostino normality test.
The gofnorm is tested with 13 methods and outputs the results.

Attached: a blog post on the Web:
Http://blog.sina.com.cn/s/blog_65efeb0c0100htz7.html
Common normal test methods for SPSS and SAS
Many measurement data analysis methods require that the data distribution is normal or approximate normal, so it is necessary to test the original independent measurement data for normality.
The data distribution normality is judged qualitatively by plotting the frequency distribution histogram of the data. Such a graphical judgment is by no means strictly a normality test, and the information provided is only an important complement to the normality test.
There are three main kinds of methods for normality test:
I. Calculation of comprehensive statistics
such as the dynamic difference method, Shapiro-Wilk Shapiro-wilk method (W test), D′agostino method (D test), Shapiro-francia method (w′ test).
Test of goodness of fit for normal distribution
such as Pearson χ2 test, log likelihood ratio test, Kolmogorov Kolmogorov-smirov method test.
Three, diagram method (normal probability diagram normal probability plot)
such as the percentile graph (Quantile quantileplot, abbreviated QQ map), percentile (Percent Percent plot, referred to as PP) and the stability of the probability map (stablized probability plot, abbreviated SP map) and so on.
Normal test methods used in statistical software
1. Test the normality of the data with the partial state coefficient and the peak state coefficient
The skewness coefficient, SK, is used to test the asymmetry, and the peak state coefficient ku, which is used to test the peak state. S k= 0, K u= 0 o'clock, the distribution is normal, s k> 0 o'clock, the distribution is positive skewness, S K < 0 o'clock, the distribution is negative bias state. Applicable conditions: The sample content should be greater than 200
2. Using the Shapiro-Wilk (Shapiro-wilk) method to test the normality of data
The W test, which was proposed in 1965, is suitable for the test of normality when the sample content is n≤50.
3. Test data normality with Dagostino (D′agostino) method
That is, D test, 1971 proposed, the normal D test the method is high efficiency, is a more accurate normal test method.
4. Shapiro-francia method
The w′ test, presented in 1972, is suitable for testing the normality of the < n < 100.
5, QQ map or PP chart
The scatter points are gathered around the fixed line, which can be considered as the approximate normal distribution of data.
Spss&sas Rules:
SPSS stipulates: When the sample content 3≤n≤5000, the results are difficult to shapiro-wilk (W test), when the sample content n > 5000 results to Kolmogorov-smirnov prevail.
SAS rules: When the sample content n≤2000, the results are Shapiro-wilk (W test), when the sample content n >2000, the results are Kolmogorov-smirnov (D test) prevail.
Reference:
Liu Qingwu, Hu Zhiyan, how to use SPSS, SAS statistical software for normality test, Journal of Shonan College (natural Science Edition), 2005
Scarlet Soldier, He Lijuan, the method of verifying the normality of data in SPSS10.0, Journal of Capital Institute of Physical Education, 2004

Test of the normality of the R language