Reply
Tue 8 Feb, 2011 08:56 pm
i have dataset on temperature and it is required that i perform a lilliefor's test on it.this will be used for fuzzy time series forecasting.I need the analytic approach with illustration.i could not get(understand) the program implementation on MATLAB.The dataset is average monthly temperature of a place in nigeria -over a period of 7 years. i need help. i am confused.
@jadeiza,
Here is a description for Lilliefor's test on MATLAB.
http://www.mathworks.com/help/toolbox/stats/lillietest.html
How large is your dataset?
@jadeiza,
From the link:
Quote:Examples
Use lillietest to determine if car mileage, in miles per gallon (MPG), follows a normal distribution across different makes of cars:
load carbig.mat
[h,p] = lillietest(MPG)
Warning: P is less than the smallest tabulated value, returning 0.001.
h =
1
p =
1.0000e-003
This is clear evidence for rejecting the null hypothesis of normality, but the p value returned is just the smallest value in the table of pre-computed values. To find a more accurate p value for the test, run a Monte Carlo approximation using the mctol input argument:
[h,p] = lillietest(MPG,0.05,'norm',1e-4)
h =
1
p =
8.3333e-006
load your data set
indicate which test format you want to run (use the Monte Carlo format if n<1000)
the program returns the test statistic and the probability of seeing that statistic by chance. p<0.05 indicates that the data are not normally distributed.