Goodness-of-fit testing for the inverse Gaussian distribution based on new entropy estimation using ranked set sampling and double ranked set sampling
- Amer Ibrahim Al-Omari^{1}Email author and
- Abdul Haq^{2}
DOI: 10.1186/2193-2697-1-8
© Al-Omari and Haq; licensee Springer. 2012
Received: 26 July 2012
Accepted: 29 August 2012
Published: 15 September 2012
Abstract
Background
Entropy is a measure of uncertainty and dispersion associated with a random variable. Several goodness-of-fit tests based on entropy are available in literature and the entropy been widely used in many applications.
Results
Goodness-of-fit test for the inverse Gaussian distribution is studied based on new entropy estimation using simple random sampling (SRS), ranked set sampling (RSS) and double ranked set sampling (DRSS) methods. The critical values of the new tests are obtained using Monte Carlo simulations. The power values of the suggested tests based on several alternative hypotheses using SRS, RSS, and DRSS are also presented. It is observed that the proposed tests are more powerful as compared to the test under SRS. Also, it turns out that the test based on DRSS is superior to the RSS test for all of the cases considered in this study.
Conclusion
Since the suggested goodness-of-fit tests for the inverse Gaussian distribution using DRSS are more efficient than that based on RSS, one may consider them using multistage RSS.
Keywords
Entropy Goodness-of-fit test Inverse Gaussian Root mean square error Simple random sampling Ranked set sampling Double ranked set samplingBackground
where m is a positive integer, known as a window size, m < n/2. Here X_{(i)} = X_{(1)} if i < 1 and X_{(i)} = X_{(1)} if i > n. It is of interest to note that$V{E}_{\left(m,n\right)}\to PH\left(f\right)$ as n → ∞, m → ∞ and m/n → 0.
They proved the consistency and asymptotic normality of this estimator under some conditions.
Based on the simulation study, it is shown that this estimator has smaller bias and mean square error as compared to the Vasicek (1976) entropy estimator. They proved that EE_{(m,n)} converges in probability to H(f) as n → ∞, m → ∞ and m/n → 0.
Alizadeh (2010) proposed a new estimator of entropy and studied its application in testing normality. Park and Park (2003) considered correcting moments for goodness-of-fit tests for two entropy estimates.
Inverse Gaussian distribution
The IG (x; μ, β) has many applications in the field, for example see Seshadri (1999), and Folks and Chhikara (1998).
Method
The test procedure
Let${X}_{1},{X}_{2},\dots ,{X}_{n}$ be a random sample of size n drawn from the pdf f(x) and let${X}_{\left(1\right)}\le {X}_{\left(2\right)}\phantom{\rule{0.5em}{0ex}}\le \cdots \le {X}_{\left(n\right)}$ be the order statistics of this sample. Our interest is to test that this random sample is coming from an inverse Gaussian population or not. Thus, the composite null hypothesis is H_{0}: X ~ IG (x; μ, β).
The following corollary is due to Mahdizaheh and Arghami (2010).
Corollary 1: Assume that X is a random variable has an inverse Gaussian distribution IG (x; μ, β) and let$Y=1/\sqrt{X}$ Then the entropy of Y is given by$H\left(f\left(y\right)\right)=log\left(0.5\phantom{\rule{0.12em}{0ex}}\varphi \sqrt{2\pi e}\right)$, where${\varphi}^{2}=1/\beta =E\left({Y}^{2}\right)-1/E\left({Y}^{-2}\right)\text{.}$
The following corollary is due to Mudholkar and Tian (2002).
Corollary 2: The random variable X with inverse Gaussian distribution IG (x; μ, β) is characterized by the property that$1/\sqrt{X}$ attains the maximum entropy among all nonnegative, absolutely continuous random variables Y with a given value at$E\left({Y}^{2}\right)-1/E\left({Y}^{-2}\right)\text{.}$
where${y}_{\left(i\right)}={\left({x}_{\left(n-i+1\right)}\right)}^{-1/2}\phantom{\rule{0.12em}{0ex}}\left(i=1,2,\dots ,n\right)\text{.}$
Suggested test
The ranked set sampling method was suggested by McIntyre (1952) for estimating the mean of pasture and forage yields. The RSS can be described as follows:
Step 1: Select n simple random samples each of size n from the target population.
Step 2: Without cost, visually rank the units within each sample with respect to the variable of interest.
Step 3: For actual measurement, from the i th$\left(i=1,2,\dots ,n\right)$ sample of n units, select the i th smallest ranked unit. The method is repeated h times if needed to increase the sample size to hn units.
Al-Saleh and Al-Kadiri (2000) suggested double ranked set sampling (DRSS) method for estimating the population mean. The DRSS can be described as in the following steps:
Step 1 Randomly select n^{2} samples each of size n from the target population.
Step 2 Apply the RSS method on the n^{2} samples obtained in Step 1. This step yields n samples each of size n.
Step 3 Reapply the RSS method again on the n samples obtained on Step 2 to obtain a sample of size n from the DRSS data. The cycle can be repeated h times if needed to obtain a sample of size hn units.
The SRS estimator of the population mean is given by${\widehat{\mu}}_{\mathit{SRS}}={\displaystyle {\sum}_{i=1}^{n}{X}_{i}}/n\text{,}$ with variance$Var\left({\widehat{\mu}}_{\mathit{SRS}}\right)={\sigma}^{2}/n$. The RSS estimator of the population mean is defined as${\widehat{\mu}}_{\mathit{RSS}}={\displaystyle {\sum}_{i=1}^{n}{{X}_{i}}_{\left(i\right)}}/n$, with variance given by$Var\left({\widehat{\mu}}_{\mathit{RSS}}\right)=\frac{{\sigma}^{2}}{n}-\frac{1}{{n}^{2}}{\displaystyle {\sum}_{i=1}^{n}{\left({\mu}_{\left(i\right)}-\mu \right)}^{2}}$. The relative precision (RP) of RSS relative to SRS for estimating the population mean is
Takahasi and Wakimoto (1968) showed that the parent pdf f (x) and the population mean can be expressed as$f\left(x\right)=\frac{1}{n}{\displaystyle {\sum}_{i=1}^{n}{f}_{\left(i\right)}\left(x\right),\phantom{\rule{0.5em}{0ex}}}and\phantom{\rule{0.5em}{0ex}}\mu =\frac{1}{n}{\displaystyle {\sum}_{i=1}^{n}{\mu}_{\left(i\right)}}$, respectively. Also, they showed that$1\le RP\le \frac{m+1}{2}$, where the lower bound is attained if and only if the underlying distribution is degenerate, while the upper bound is attained if and only if the underlying distribution of the data is rectangular.
Al-Saleh and Al-Omari (2002) extended the DRSS for multistage RSS method to increase the efficiency of the estimators for fixed value of the sample size, Al-Omari and Raqab (2012) suggested truncation RSS method for estimating the population mean and median, Al-Omari (2011) suggested double robust extreme RSS for estimating the population mean, Haq and Shabbir (2010) proposed a family of ratio estimators of the population mean using extreme RSS based on two auxiliary variables.
where
Note that,$A{E}_{\left(m,n\right)}\left({f}_{y}\right)$ is the sample estimate of$AE\left({f}_{y}\right)$. Since the entropy estimators are functions of order statistics, then the entropy estimation using RSS and DRSS involves ordering the RSS units.
Results and discussion
In this section, a Monte Carlo experiment is presented to investigate the performance of the entropy estimators i.e. AE_{(m,n)} as well as VE_{(m,n)} and as well as to study the powers of the suggested tests under different alternatives hypotheses. The root mean square errors (RMSEs) and the bias values are obtained for the estimators based on 10,000 samples of sizes n = 10, 20, 30 with window sizes 1 ≤ m ≤5, 1 ≤ m ≤10 and 1 ≤ m ≤ 15, respectively.
Comparison between VE_{(m,n)}and AE_{(m,n)}
Monte Carlo RMSEs and bias values of the entropy estimators VE_{ ( m,n ) } and AE_{ ( m,n ) } for the uniform distribution , H(f) = 0
n | m | SRS | RSS | ||||||
---|---|---|---|---|---|---|---|---|---|
VE _{(m,n)} | AE _{(m,n)} | VE _{(m,n)} | AE _{(m,n)} | ||||||
Bias | RMSE | Bias | RMSE | Bias | RMSE | Bias | RMSE | ||
10 | 1 | −0.519826 | 0.569537 | −0.046482 | 0.521035 | −0.396308 | 0.443439 | −0.343522 | 0.396739 |
2 | −0.415135 | 0.452358 | −0.298609 | 0.350332 | −0.304078 | 0.329233 | −0.189664 | 0.228762 | |
3 | −0.422613 | 0.453818 | −0.249056 | 0.298944 | −0.327681 | 0.343991 | −0.154894 | 0.186380 | |
4 | −0.458940 | 0.487054 | −0.229082 | 0.281422 | −0.371538 | 0.383103 | −0.143218 | 0.171767 | |
5 | −0.502063 | 0.527918 | −0.215077 | 0.270468 | −0.425903 | 0.436521 | −0.137821 | 0.168029 | |
20 | 1 | −0.393900 | 0.418346 | −0.366867 | 0.392622 | −0.343340 | 0.365754 | −0.314244 | 0.338695 |
2 | −0.271880 | 0.290818 | −0.212993 | 0.236696 | −0.217937 | 0.233026 | −0.162729 | 0.183187 | |
3 | −0.253931 | 0.270200 | −0.168961 | 0.192998 | −0.205321 | 0.216879 | −0.117939 | 0.136570 | |
4 | −0.260596 | 0.274678 | −0.144016 | 0.167779 | −0.214042 | 0.222524 | −0.100304 | 0.118284 | |
5 | −0.276800 | 0.288985 | −0.133179 | 0.157805 | −0.235141 | 0.242179 | −0.091608 | 0.108584 | |
6 | −0.299321 | 0.310256 | −0.125960 | 0.150733 | −0.258899 | 0.264554 | −0.085981 | 0.101365 | |
7 | −0.322084 | 0.332301 | −0.121244 | 0.146386 | −0.285310 | 0.290156 | −0.084733 | 0.099613 | |
8 | −0.348254 | 0.357901 | −0.118562 | 0.144786 | −0.314138 | 0.318471 | −0.083482 | 0.098588 | |
9 | −0.374620 | 0.383864 | −0.116399 | 0.143986 | −0.343410 | 0.347711 | −0.083926 | 0.099430 | |
10 | −0.402840 | 0.411741 | −0.117057 | 0.145063 | −0.371780 | 0.375737 | −0.848235 | 0.101014 | |
30 | 1 | −0.352853 | 0.368369 | −0.334631 | 0.351096 | −0.319230 | 0.333509 | −0.300423 | 0.316118 |
2 | −0.223356 | 0.235685 | −0.184969 | 0.199765 | −0.190866 | 0.201625 | −0.152577 | 0.165665 | |
3 | −0.197719 | 0.208362 | −0.141411 | 0.156683 | −0.165182 | 0.173360 | −0.106329 | 0.119047 | |
4 | −0.196240 | 0.205882 | −0.118803 | 0.133958 | −0.162899 | 0.169841 | −0.087046 | 0.099566 | |
5 | −0.202003 | 0.210395 | −0.105711 | 0.120861 | −0.172441 | 0.178293 | −0.078599 | 0.088072 | |
6 | −0.213804 | 0.221385 | −0.097719 | 0.113216 | −0.185622 | 0.190458 | −0.069898 | 0.081972 | |
7 | −0.226688 | 0.233521 | −0.092957 | 0.109089 | −0.200036 | 0.204048 | −0.066053 | 0.077716 | |
8 | −0.242599 | 0.248992 | −0.089259 | 0.105818 | −0.217704 | 0.221309 | −0.064713 | 0.076188 | |
9 | −0.259471 | 0.265356 | −0.087074 | 0.103535 | −0.235661 | 0.238850 | −0.062931 | 0.073734 | |
10 | −0.276934 | 0.282548 | −0.085151 | 0.102071 | −0.254437 | 0.257257 | −0.062044 | 0.072402 | |
11 | −0.295302 | 0.300725 | −0.841357 | 0.101314 | −0.273700 | 0.276336 | −0.062243 | 0.072977 | |
12 | −0.313803 | 0.319255 | −0.083206 | 0.102002 | −0.293398 | 0.295911 | −0.062262 | 0.072981 | |
13 | −0.332279 | 0.337432 | −0.082858 | 0.101944 | −0.311978 | 0.341101 | −0.063754 | 0.074987 | |
14 | −0.351090 | 0.356205 | −0.082540 | 0.101854 | −0.332096 | 0.334518 | −0.063579 | 0.075100 | |
15 | −0.370555 | 0.375518 | −0.082665 | 0.102618 | −0.352077 | 0.354327 | −0.064127 | 0.075825 |
Monte Carlo RMSEs and bias values of the entropy estimators VE_{ ( m,n ) } and AE_{ ( m,n ) } for the exponential distribution, H(f) = 1
n | m | SRS | RSS | ||||||
---|---|---|---|---|---|---|---|---|---|
VE _{(m,n)} | AE _{(m,n)} | VE _{(m,n)} | AE _{(m,n)} | ||||||
Bias | RMSE | Bias | RMSE | Bias | RMSE | Bias | RMSE | ||
10 | 1 | −0.552032 | 0.677001 | −0.495449 | 0.631471 | −0.430553 | 0.505229 | −0.376361 | 0.461201 |
2 | −0.442683 | 0.571820 | −0.323532 | 0.483573 | −0.337494 | 0.404667 | −0.220406 | 0.315220 | |
3 | −0.435444 | 0.561640 | −0.265713 | 0.443276 | −0.332760 | 0.401125 | −0.159787 | 0.276197 | |
4 | −0.451545 | 0.575390 | −0.221689 | 0.424404 | −0.348029 | 0.420617 | −0.121584 | 0.266664 | |
5 | −0.469437 | 0.597761 | −0.179844 | 0.413541 | −0.366628 | 0.445977 | −0.080667 | 0.266812 | |
20 | 1 | −0.414064 | 0.490107 | −0.384516 | 0.464796 | −0.357765 | 0.398661 | −0.333540 | 0.376843 |
2 | −0.285717 | 0.376086 | −0.232518 | 0.338830 | −0.234959 | 0.280262 | −0.176512 | 0.232710 | |
3 | −0.260773 | 0.351341 | −0.175461 | 0.298406 | −0.213397 | 0.261261 | −0.125059 | 0.194705 | |
4 | −0.256116 | 0.352810 | 0.141143 | 0.279706 | −0.210620 | 0.259248 | −0.098056 | 0.179990 | |
5 | −0.262412 | 0.358638 | 0.118697 | 0.271887 | −0.214122 | 0.265246 | −0.072456 | 0.172661 | |
6 | −0.265650 | 0.360325 | 0.090043 | 0.263318 | −0.218028 | 0.272315 | −0.048075 | 0.168086 | |
7 | −0.266934 | 0.365008 | −0.067175 | 0.260090 | −0.224596 | 0.282196 | −0.023128 | 0.173677 | |
8 | −0.273952 | 0.377519 | −0.041928 | 0.258647 | −0.232629 | 0.293062 | −0.000531 | 0.176806 | |
9 | −0.280123 | 0.381968 | −0.021108 | 0.262708 | −0.236125 | 0.302083 | 0.027269 | 0.190739 | |
10 | −0.285183 | 0.391290 | 0.004497 | 0.267634 | −0.238413 | 0.310922 | 0.044912 | 0.203657 | |
30 | 1 | −0.367058 | 0.423423 | −0.346283 | 0.406311 | −0.332526 | 0.361491 | −0.313657 | 0.343784 |
2 | −0.233677 | 0.306086 | −0.198867 | 0.280012 | −0.203455 | 0.236001 | −0.163180 | 0.203230 | |
3 | −0.202277 | 0.281503 | −0.145618 | 0.241162 | −0.170859 | 0.207468 | −0.111717 | 0.161754 | |
4 | −0.194424 | 0.275072 | −0.115163 | 0.224526 | −0.160246 | 0.199410 | −0.084854 | 0.145930 | |
5 | −0.191705 | 0.272356 | −0.095073 | 0.217468 | −0.159714 | 0.200465 | −0.059819 | 0.134539 | |
6 | −0.186870 | 0.272196 | −0.070590 | 0.208597 | −0.158702 | 0.202869 | −0.043778 | 0.132887 | |
7 | −0.191094 | 0.275374 | −0.058550 | 0.205261 | −0.161705 | 0.206226 | −0.027194 | 0.130283 | |
8 | −0.195662 | 0.280589 | −0.036080 | 0.200329 | −0.164468 | 0.212265 | −0.010631 | 0.136358 | |
9 | −0.196983 | 0.282040 | −0.021144 | 0.202056 | −0.165511 | 0.217222 | −0.006685 | 0.138626 | |
10 | −0.197171 | 0.283394 | −0.005890 | 0.204787 | −0.167152 | 0.220237 | 0.024904 | 0.145306 | |
11 | −0.198853 | 0.286241 | 0.008492 | 0.207709 | −0.173076 | 0.229318 | 0.039837 | 0.154215 | |
12 | −0.204089 | 0.293653 | 0.022622 | 0.213445 | −0.171555 | 0.232740 | 0.055108 | 0.163320 | |
13 | −0.202908 | 0.298108 | 0.049154 | 0.220522 | −0.176996 | 0.240454 | 0.070977 | 0.176787 | |
14 | −0.205700 | 0.300842 | 0.061987 | 0.226574 | −0.176922 | 0.244541 | 0.093001 | 0.193377 | |
15 | −0.210699 | 0.305809 | 0.081431 | 0.238902 | −0.177959 | 0.248760 | 0.109754 | 0.205539 |
Monte Carlo RMSEs and bias values of the entropy estimators VE_{ ( m,n ) } and AE_{ ( m,n ) } for the standard normal distribution, H(f) = 1.419
n | m | SRS | RSS | ||||||
---|---|---|---|---|---|---|---|---|---|
VE _{(m,n)} | AE _{(m,n)} | VE _{(m,n)} | AE _{(m,n)} | ||||||
Bias | RMSE | Bias | RMSE | Bias | RMSE | Bias | RMSE | ||
10 | 1 | −0.598925 | 0.676499 | −0.538428 | 0.623068 | −0.484489 | 0.549750 | −0.429406 | 0.502967 |
2 | −0.521455 | 0.591007 | −0.409842 | 0.496627 | −0.422169 | 0.471157 | −0.308706 | 0.375690 | |
3 | −0.563002 | 0.623188 | −0.386562 | 0.468471 | −0.462240 | 0.504378 | −0.291133 | 0.353844 | |
4 | −0.610651 | 0.663364 | 0.388846 | 0.469519 | −0.523019 | 0.557792 | −0.292810 | 0.351636 | |
5 | −0.671777 | 0.719069 | −0.382242 | 0.461612 | −0.584483 | 0.614209 | −0.294820 | 0.349472 | |
20 | 1 | −0.435480 | 0.483459 | −0.402721 | 0.452976 | −0.382986 | 0.420310 | −0.354315 | 0.393878 |
2 | −0.327145 | 0.375798 | −0.267005 | 0.324501 | −0.275716 | 0.313472 | −0.218758 | 0.264068 | |
3 | −0.317948 | 0.364927 | −0.230598 | 0.292997 | −0.268657 | 0.304811 | −0.181588 | 0.230636 | |
4 | −0.327070 | 0.372436 | −0.214227 | 0.279269 | −0.285331 | 0.318855 | −0.168035 | 0.219922 | |
5 | −0.352658 | 0.395796 | −0.205782 | 0.272804 | −0.305555 | 0.337744 | −0.160392 | 0.213700 | |
6 | 0.375996 | 0.416964 | −0.203268 | 0.269194 | −0.335066 | 0.365185 | −0.162263 | 0.216405 | |
7 | −0.404050 | 0.442997 | −0.200951 | 0.269828 | −0.363782 | 0.391748 | −0.162648 | 0.217866 | |
8 | −0.439618 | 0.475094 | −0.203704 | 0.270603 | −0.395221 | 0.421583 | −0.163443 | 0.217711 | |
9 | −0.467134 | 0.500777 | 0.211872 | 0.276695 | −0.428042 | 0.451680 | −0.169841 | 0.224475 | |
10 | −0.496926 | 0.527456 | −0.209085 | 0.275281 | −0.454818 | 0.477152 | −0.171572 | 0.224804 | |
30 | 1 | −0.378860 | 0.413455 | −0.359097 | 0.394766 | −0.343626 | 0.370512 | −0.328056 | 0.355718 |
2 | −0.259105 | 0.299687 | −0.221750 | 0.266138 | −0.226914 | 0.255947 | −0.189446 | 0.223276 | |
3 | −0.236758 | 0.277238 | −0.177599 | 0.229027 | −0.204698 | 0.234358 | −0.147274 | 0.186797 | |
4 | −0.234369 | 0.275867 | −0.158560 | 0.213972 | −0.204765 | 0.234413 | −0.125487 | 0.169031 | |
5 | −0.244288 | 0.283027 | −0.148610 | 0.206988 | −0.214434 | 0.243683 | −0.117590 | 0.165087 | |
6 | −0.255248 | 0.293332 | −0.139542 | 0.200072 | −0.227340 | 0.255901 | −0.111407 | 0.161770 | |
7 | −0.269724 | 0.305134 | −0.132038 | 0.196792 | −0.241325 | 0.268228 | −0.105796 | 0.158654 | |
8 | −0.285713 | 0.321039 | −0.129915 | 0.193509 | −0.254983 | 0.282376 | −0.102504 | 0.157726 | |
9 | −0.304064 | 0.337563 | −0.131105 | 0.198239 | −0.274697 | 0.301420 | −0.103392 | 0.160749 | |
10 | −0.320051 | 0.352764 | −0.130086 | 0.196928 | −0.295057 | 0.319933 | −0.101392 | 0.160593 | |
11 | −0.339131 | 0.369866 | −0.127890 | 0.196985 | −0.314201 | 0.339141 | −0.102034 | 0.161378 | |
12 | −0.361226 | 0.392070 | −0.130212 | 0.197655 | −0.333173 | 0.356224 | −0.103026 | 0.163577 | |
13 | −0.382347 | 0.410463 | 0.129885 | 0.199488 | −0.353582 | 0.375170 | −0.105978 | 0.165825 | |
14 | −0.400618 | 0.428008 | −0.131518 | 0.199794 | −0.375752 | 0.397462 | −0.109190 | 0.168154 | |
15 | −0.423597 | 0.449968 | −0.134062 | 0.200285 | −0.394363 | 0.414605 | −0.108705 | 0.167780 |
Monte Carlo RMSEs and bias values of the entropy estimators VE_{ ( m,n ) } and AE_{ ( m,n ) } for the uniform distribution with H(f) = 0 and exponential distribution with H(f) = 1 using DRSS
n | m | Uniform distribution and$H\left(f\right)=0$ | Exponential distribution and$H\left(f\right)=1$ | ||||||
---|---|---|---|---|---|---|---|---|---|
VE _{(m,n)} | AE _{(m,n)} | VE _{(m,n)} | AE _{(m,n)} | ||||||
Bias | RMSE | Bias | RMSE | Bias | RMSE | Bias | RMSE | ||
10 | 1 | −0.327408 | 0.369593 | −0.267924 | 0.318205 | −0.365854 | 0.425279 | −0.305667 | 0.379121 |
2 | −0.260621 | 0.278731 | −0.145388 | 0.176159 | −0.288898 | 0.340618 | −0.173991 | 0.251460 | |
3 | −0.296104 | 0.306116 | −0.122180 | 0.144286 | −0.300393 | 0.351750 | −0.128545 | 0.223802 | |
4 | −0.346305 | 0.352712 | −0.115995 | 0.134276 | −0.322839 | 0.377437 | −0.089495 | 0.215854 | |
5 | −0.404121 | 0.409902 | −0.116805 | 0.135411 | −0.335248 | 0.399189 | −0.047170 | 0.219634 | |
20 | 1 | −0.308453 | 0.329353 | −0.279902 | 0.302719 | −0.329105 | 0.363241 | −0.298237 | 0.335475 |
2 | −0.189231 | 0.202666 | −0.132076 | 0.151177 | −0.204908 | 0.240316 | −0.150759 | 0.196279 | |
3 | −0.182095 | 0.191163 | −0.095961 | 0.112229 | −0.191216 | 0.228320 | −0.104346 | 0.163293 | |
4 | −0.197693 | 0.204342 | −0.082268 | 0.096978 | −0.190904 | 0.229986 | −0.075338 | 0.179771 | |
5 | −0.220876 | 0.225845 | −0.077708 | 0.091093 | −0.197900 | 0.239789 | −0.052175 | 0.145269 | |
6 | −0.247733 | 0.251580 | −0.075071 | 0.086966 | −0.207032 | 0.251002 | −0.026183 | 0.146832 | |
7 | −0.275808 | 0.278919 | −0.074331 | 0.085055 | −0.209883 | 0.258152 | −0.012044 | 0.152682 | |
8 | −0.303823 | 0.306608 | −0.073793 | 0.084202 | −0.218701 | 0.271560 | 0.014201 | 0.161180 | |
9 | −0.333903 | 0.336495 | −0.075306 | 0.086127 | −0.223692 | 0.278728 | 0.035069 | 0.173654 | |
10 | −0.363272 | 0.365731 | −0.075514 | 0.086480 | −0.228126 | 0.290431 | 0.061574 | 0.189857 | |
30 | 1 | −0.298092 | 0.312767 | −0.278830 | 0.293698 | −0.308011 | 0.331033 | −0.289677 | 0.314515 |
2 | −0.170745 | 0.180210 | −0.133715 | 0.146379 | −0.182416 | 0.207785 | −0.143418 | 0.174632 | |
3 | −0.146113 | 0.153646 | −0.088998 | 0.100564 | −0.152039 | 0.180708 | −0.094799 | 0.136371 | |
4 | −0.149143 | 0.154886 | −0.072297 | 0.083848 | −0.145325 | 0.176699 | −0.071094 | 0.123270 | |
5 | −0.159888 | 0.164564 | −0.063874 | 0.074562 | −0.146632 | 0.179028 | −0.049250 | 0.114227 | |
6 | −0.174419 | 0.178204 | −0.060394 | 0.070784 | −0.149443 | 0.184598 | −0.030887 | 0.113500 | |
7 | −0.191854 | 0.194940 | −0.058041 | 0.067650 | −0.150245 | 0.188158 | −0.046556 | 0.115023 | |
8 | −0.209886 | 0.212509 | −0.056421 | 0.065369 | −0.153441 | 0.194332 | −0.001239 | 0.120306 | |
9 | −0.229010 | 0.231261 | −0.056053 | 0.064628 | −0.157250 | 0.199936 | 0.012716 | 0.124585 | |
10 | −0.248006 | 0.249993 | −0.056843 | 0.064868 | −0.162854 | 0.208891 | 0.029477 | 0.133242 | |
11 | −0.267506 | 0.269188 | −0.056931 | 0.064430 | −0.163540 | 0.213175 | 0.045951 | 0.145582 | |
12 | −0.287408 | 0.289018 | −0.056982 | 0.064673 | −0.167660 | 0.221482 | 0.063602 | 0.155340 | |
13 | −0.307160 | 0.308699 | −0.058363 | 0.066130 | −0.171024 | 0.225764 | 0.079779 | 0.169499 | |
14 | −0.327370 | 0.328890 | −0.058038 | 0.065797 | −0.170880 | 0.232977 | 0.096359 | 0.182124 | |
15 | −0.346997 | 0.348439 | −0.059523 | 0.067623 | −0.169873 | 0.235173 | 0.115563 | 0.198755 |
Monte Carlo RMSEs and bias values of the entropy estimators VE_{ ( m,n ) } and AE_{ ( m,n ) } for the standard normal distribution and H(f) = 1.419 using DRSS
n | m | VE _{(m,n)} | AE _{(m,n)} | ||
---|---|---|---|---|---|
Bias | RMSE | Bias | RMSE | ||
10 | 1 | −0.415021 | 0.472162 | −0.352434 | 0.416211 |
2 | −0.373395 | 0.412666 | −0.262149 | 0.316029 | |
3 | −0.427401 | 0.459119 | −0.254450 | 0.303820 | |
4 | −0.492911 | 0.518275 | −0.264683 | 0.310442 | |
5 | −0.554351 | 0.577281 | −0.267798 | 0.312339 | |
20 | 1 | −0.350703 | 0.383160 | −0.323780 | 0.359592 |
2 | −0.245907 | 0.277809 | −0.190733 | 0.231106 | |
3 | −0.246496 | 0.276941 | −0.158832 | 0.201924 | |
4 | −0.262789 | 0.290545 | −0.148107 | 0.194728 | |
5 | −0.291340 | 0.317967 | −0.145734 | 0.191755 | |
6 | −0.316105 | 0.341597 | −0.147800 | 0.195946 | |
7 | −0.349246 | 0.373132 | −0.150312 | 0.199934 | |
8 | −0.384526 | 0.406764 | −0.152801 | 0.203493 | |
9 | −0.416151 | 0.436696 | −0.156902 | 0.205954 | |
10 | −0.445901 | 0.465518 | 0.159050 | 0.207883 | |
30 | 1 | −0.321940 | 0.345223 | −0.307781 | 0.332609 |
2 | −0.206709 | 0.231560 | −0.169564 | 0.198438 | |
3 | −0.187163 | 0.212774 | −0.129694 | 0.163913 | |
4 | −0.190073 | 0.215577 | −0.114103 | 0.152713 | |
5 | −0.199843 | 0.224569 | −0.103570 | 0.145964 | |
6 | −0.214636 | 0.239021 | −0.100510 | 0.146417 | |
7 | −0.231613 | 0.255278 | −0.095517 | 0.143483 | |
8 | −0.247340 | 0.271084 | −0.094560 | 0.145579 | |
9 | −0.268298 | 0.291044 | −0.091548 | 0.145394 | |
10 | −0.286538 | 0.308661 | −0.094236 | 0.149024 | |
11 | −0.305310 | 0.326485 | −0.093843 | 0.150300 | |
12 | −0.324892 | 0.346062 | −0.096171 | 0.152896 | |
13 | −0.343097 | 0.363236 | −0.096892 | 0.153854 | |
14 | −0.369990 | 0.388586 | −0.100541 | 0.155029 | |
15 | −0.387740 | 0.406081 | −0.101202 | 0.156143 |
Critical values of the test statistics at significance level α = 0.05 using SRS, RSS and DRSS
n = 30 | ||||||||
---|---|---|---|---|---|---|---|---|
n | m | SRS | RSS | DRSS | m | SRS | RSS | DRSS |
10 | 1 | 1.77481 | 1.92014 | 2.11693 | 1 | 2.45932 | 2.50879 | 2.57507 |
2 | 2.32375 | 2.49737 | 2.73051 | 2 | 3.00586 | 3.06976 | 3.15363 | |
3 | 2.55582 | 2.70474 | 2.87862 | 3 | 3.19857 | 3.25881 | 3.33729 | |
4 | 2.67573 | 2.81527 | 2.91803 | 4 | 3.27582 | 3.35586 | 3.42156 | |
5 | 2.73289 | 2.83557 | 2.91884 | 5 | 3.32359 | 3.39547 | 3.45623 | |
20 | 1 | 2.24771 | 2.35314 | 2.42654 | 6 | 3.35015 | 3.42129 | 3.47623 |
2 | 2.79602 | 2.88869 | 3.02510 | 7 | 3.36693 | 3.43050 | 3.47907 | |
3 | 2.97493 | 3.08786 | 3.19524 | 8 | 3.37529 | 3.43391 | 3.47352 | |
4 | 3.04798 | 3.15706 | 3.25697 | 9 | 3.37021 | 3.43604 | 3.47057 | |
5 | 3.09802 | 3.19645 | 3.28312 | 10 | 3.38831 | 3.43064 | 3.47215 | |
6 | 3.13033 | 3.21615 | 3.28262 | 11 | 3.39279 | 3.42939 | 3.45317 | |
7 | 3.15950 | 3.22789 | 3.27655 | 12 | 3.38330 | 3.41772 | 3.44495 | |
8 | 3.15719 | 3.21777 | 3.26882 | 13 | 3.37597 | 3.42184 | 3.44197 | |
9 | 3.16680 | 3.21856 | 3.26432 | 14 | 3.36220 | 3.41612 | 3.44014 | |
10 | 3.15824 | 3.21474 | 3.25051 | 15 | 3.38366 | 3.41508 | 3.43684 |
Optimal window sizes
n | SRS | RSS | DRSS |
---|---|---|---|
10 | 5 | 5 | 5 |
20 | 9 | 7 | 5 |
30 | 11 | 9 | 7 |
We can see that these optimal values are different from Mahdizaheh and Arghami (2010) values where their suggested test is based on Vasicek (1976) entropy estimator. Here, we can conclude that the optimal window size depends on the entropy estimator used for the goodness-of-fit test.
Power of the tests
The power of the suggested goodness-of-fit tests using SRS, RSS and DRSS is considered here relative to the same alternatives considered by Mahdizaheh and Arghami (2010) for the distributions, exponential(1), uniform(0,1), Weibull(2,1), lognormal(0,2), beta(2,2), and beta(5,2). 10000 samples of sizes n = 30, 20, 30 are generated for each method at the significance level 0.05.
Power comparison for the entropy tests at the significance level α = 0.05
n | m | Exponential (1) | Uniform (0,1) | Weibull (2,1) | ||||||
---|---|---|---|---|---|---|---|---|---|---|
SRS | RSS | DRSS | SRS | RSS | DRSS | SRS | RSS | DRSS | ||
10 | 1 | 0.1869 | 0.2330 | 0.2559 | 0.4089 | 0.5078 | 0.5921 | 0.1059 | 0.1238 | 0.1346 |
2 | 0.2167 | 0.2776 | 0.3610 | 0.4874 | 0.6422 | 0.8381 | 0.1269 | 0.1640 | 0.2240 | |
3 | 0.1960 | 0.2562 | 0.3242 | 0.4796 | 0.6398 | 0.8455 | 0.1261 | 0.1659 | 0.2230 | |
4 | 0.1366 | 0.1875 | 0.1981 | 0.3735 | 0.5284 | 0.6825 | 0.0961 | 0.1391 | 0.1593 | |
5 | 0.0629 | 0.0750 | 0.0780 | 0.1897 | 0.2481 | 0.3011 | 0.0460 | 0.0574 | 0.0622 | |
20 | 1 | 0.3805 | 0.4530 | 0.4682 | 0.7665 | 0.8704 | 0.9186 | 0.1874 | 0.2311 | 0.2351 |
2 | 0.4584 | 0.5375 | 0.6152 | 0.8661 | 0.9528 | 0.9930 | 0.2566 | 0.3062 | 0.3597 | |
3 | 0.4713 | 0.5680 | 0.6360 | 0.8873 | 0.9716 | 0.9970 | 0.2625 | 0.3341 | 0.3890 | |
4 | 0.4179 | 0.5201 | 0.6027 | 0.8711 | 0.9680 | 0.9968 | 0.2299 | 0.2964 | 0.3552 | |
5 | 0.3829 | 0.4685 | 0.5284 | 0.8346 | 0.9484 | 0.9944 | 0.2095 | 0.2648 | 0.3106 | |
6 | 0.3094 | 0.3855 | 0.4221 | 0.8024 | 0.9211 | 0.9802 | 0.1682 | 0.2106 | 0.2364 | |
7 | 0.2377 | 0.2899 | 0.3074 | 0.7229 | 0.8564 | 0.9312 | 0.1368 | 0.1611 | 0.1635 | |
8 | 0.1660 | 0.1827 | 0.1942 | 0.5806 | 0.7019 | 0.7954 | 0.0877 | 0.0955 | 0.0963 | |
9 | 0.1022 | 0.1131 | 0.1132 | 0.4095 | 0.4875 | 0.5456 | 0.0600 | 0.0581 | 0.0633 | |
10 | 0.0538 | 0.0615 | 0.0638 | 0.2145 | 0.2585 | 0.2627 | 0.0297 | 0.0328 | 0.0346 | |
30 | 1 | 0.5400 | 0.5913 | 0.6094 | 0.9188 | 0.9660 | 0.9851 | 0.2729 | 0.3091 | 0.3125 |
2 | 0.6402 | 0.7097 | 0.7585 | 0.9724 | 0.9960 | 0.9997 | 0.3776 | 0.4276 | 0.4669 | |
3 | 0.6734 | 0.7431 | 0.7941 | 0.9832 | 0.9982 | 0.9999 | 0.4116 | 0.4605 | 0.5075 | |
4 | 0.6510 | 0.7374 | 0.7959 | 0.9804 | 0.9989 | 1.0000 | 0.3941 | 0.4650 | 0.5156 | |
5 | 0.6252 | 0.7048 | 0.7711 | 0.9800 | 0.9979 | 0.9999 | 0.3636 | 0.4324 | 0.4829 | |
6 | 0.5763 | 0.6583 | 0.7229 | 0.9690 | 0.9978 | 0.9998 | 0.3109 | 0.3757 | 0.4322 | |
7 | 0.5170 | 0.6015 | 0.6531 | 0.9558 | 0.9940 | 0.9995 | 0.2795 | 0.3274 | 0.3575 | |
8 | 0.4526 | 0.5237 | 0.5565 | 0.9392 | 0.9875 | 0.9982 | 0.2166 | 0.2672 | 0.2778 | |
9 | 0.3843 | 0.4356 | 0.4609 | 0.8973 | 0.9730 | 0.9949 | 0.1768 | 0.2066 | 0.2134 | |
10 | 0.3102 | 0.3424 | 0.3547 | 0.8673 | 0.9445 | 0.9823 | 0.1421 | 0.1438 | 0.1592 | |
11 | 0.2440 | 0.2528 | 0.2585 | 0.7882 | 0.8763 | 0.9285 | 0.1066 | 0.1070 | 0.1020 | |
12 | 0.1772 | 0.1788 | 0.1785 | 0.6678 | 0.7474 | 0.8160 | 0.0713 | 0.0697 | 0.0660 | |
13 | 0.1117 | 0.1218 | 0.1141 | 0.5201 | 0.6034 | 0.6372 | 0.0447 | 0.0501 | 0.0502 | |
14 | 0.0697 | 0.0774 | 0.0800 | 0.3516 | 0.4083 | 0.4327 | 0.0269 | 0.0363 | 0.0288 | |
15 | 0.0477 | 0.0448 | 0.0522 | 0.2284 | 0.2458 | 0.2411 | 0.0231 | 0.0261 | 0.0197 |
Power comparison for the entropy tests at the significance level α = 0.05
n | m | Lognormal (0,2) | Beta (2,2) | Beta (5,2) | ||||||
---|---|---|---|---|---|---|---|---|---|---|
SRS | RSS | DRSS | SRS | RSS | DRSS | SRS | RSS | DRSS | ||
10 | 1 | 0.1347 | 0.1595 | 0.1806 | 0.1758 | 0.1990 | 0.2343 | 0.1436 | 0.1667 | 0.1823 |
2 | 0.1576 | 0.1849 | 0.2383 | 0.2208 | 0.2925 | 0.4210 | 0.2027 | 0.2649 | 0.3855 | |
3 | 0.1177 | 0.1532 | 0.1853 | 0.2341 | 0.3255 | 0.4670 | 0.2443 | 0.3276 | 0.5106 | |
4 | 0.0667 | 0.0894 | 0.0936 | 0.1871 | 0.2774 | 0.3626 | 0.2303 | 0.3554 | 0.4872 | |
5 | 0.0262 | 0.0267 | 0.0241 | 0.0910 | 0.1194 | 0.1480 | 0.1644 | 0.2462 | 0.3241 | |
20 | 1 | 0.2802 | 0.3461 | 0.3535 | 0.3543 | 0.4343 | 0.4556 | 0.2923 | 0.3556 | 0.3693 |
2 | 0.3447 | 0.4144 | 0.4731 | 0.4954 | 0.5982 | 0.7032 | 0.4418 | 0.5150 | 0.6393 | |
3 | 0.3504 | 0.4282 | 0.4726 | 0.5214 | 0.6633 | 0.7879 | 0.4817 | 0.6162 | 0.7499 | |
4 | 0.3037 | 0.3743 | 0.4325 | 0.5056 | 0.6472 | 0.7819 | 0.4799 | 0.6238 | 0.7869 | |
5 | 0.2402 | 0.3071 | 0.3379 | 0.4875 | 0.6170 | 0.7554 | 0.4742 | 0.6288 | 0.7809 | |
6 | 0.1870 | 0.2164 | 0.2338 | 0.4256 | 0.5471 | 0.6569 | 0.4546 | 0.5935 | 0.7156 | |
7 | 0.1251 | 0.1346 | 0.1326 | 0.3672 | 0.4858 | 0.5137 | 0.4299 | 0.5452 | 0.6399 | |
8 | 0.0669 | 0.0671 | 0.0720 | 0.2603 | 0.3153 | 0.3578 | 0.3735 | 0.4543 | 0.5274 | |
9 | 0.0324 | 0.0317 | 0.0323 | 0.1594 | 0.1886 | 0.2044 | 0.3094 | 0.3651 | 0.4164 | |
10 | 0.0116 | 0.0126 | 0.0136 | 0.0868 | 0.0973 | 0.0967 | 0.2227 | 0.2661 | 0.2867 | |
30 | 1 | 0.4096 | 0.4578 | 0.4737 | 0.5287 | 0.5856 | 0.6167 | 0.4344 | 0.4767 | 0.5121 |
2 | 0.5141 | 0.5748 | 0.6309 | 0.7055 | 0.7838 | 0.8603 | 0.6237 | 0.7156 | 0.7936 | |
3 | 0.5292 | 0.6032 | 0.6622 | 0.7543 | 0.8437 | 0.9182 | 0.6911 | 0.7996 | 0.8857 | |
4 | 0.5187 | 0.6013 | 0.6542 | 0.7542 | 0.8670 | 0.9382 | 0.6993 | 0.8376 | 0.9258 | |
5 | 0.4831 | 0.5571 | 0.5990 | 0.7308 | 0.8530 | 0.9339 | 0.7030 | 0.8398 | 0.9240 | |
6 | 0.4209 | 0.4965 | 0.5441 | 0.7038 | 0.8338 | 0.9185 | 0.6877 | 0.8228 | 0.9141 | |
7 | 0.3574 | 0.4220 | 0.4439 | 0.6584 | 0.7854 | 0.8702 | 0.6559 | 0.7989 | 0.8874 | |
8 | 0.2916 | 0.3275 | 0.3447 | 0.5932 | 0.7100 | 0.7995 | 0.6239 | 0.7564 | 0.8375 | |
9 | 0.2172 | 0.2460 | 0.2466 | 0.5197 | 0.6383 | 0.7055 | 0.5672 | 0.7001 | 0.7779 | |
10 | 0.1442 | 0.1705 | 0.1664 | 0.4502 | 0.5295 | 0.5999 | 0.5433 | 0.6273 | 0.7271 | |
11 | 0.1055 | 0.1037 | 0.0977 | 0.3810 | 0.4140 | 0.4532 | 0.4848 | 0.5615 | 0.6114 | |
12 | 0.0549 | 0.0555 | 0.0599 | 0.2764 | 0.2975 | 0.3117 | 0.4196 | 0.4751 | 0.5126 | |
13 | 0.0311 | 0.0288 | 0.0285 | 0.1922 | 0.2188 | 0.2187 | 0.3449 | 0.4049 | 0.4171 | |
14 | 0.0129 | 0.0148 | 0.0148 | 0.1130 | 0.1356 | 0.1376 | 0.2720 | 0.3261 | 0.3560 | |
15 | 0.0067 | 0.0070 | 0.0070 | 0.0824 | 0.0830 | 0.0822 | 0.2466 | 0.2687 | 0.2729 |
Conclusion
In this paper, new goodness-of-fit tests for the inverse Gaussian distribution are suggested using SRS, RSS and DRSS based on the maximum entropy characterization. It is found that the new tests are more powerful under RSS and DRSS, and the test under DRSS is superior to the tests under RSS and SRS methods. We recommend using the suggested goodness-of-fit tests for the inverse Gaussian distribution. As the DRSS is better than RSS, the current work can be extended to multistage RSS design and for some other probability distributions.
Declarations
Acknowledgment
The authors are grateful to the editors and the anonymous reviewers for their valuable comments and suggestions.
Authors’ Affiliations
References
- Alizadeh HN: A new estimator of entropy and its application in testing normality. J Stat Comput Simul 2010, 80: 1151–1162. 10.1080/00949650903005656View ArticleGoogle Scholar
- Al-Omari AI: Estimation of mean based on modified robust extreme ranked set sampling. J Stat Comput Simul 2011,81(8):1055–1066. 10.1080/00949651003649161View ArticleGoogle Scholar
- Al-Omari AI, Raqab MZ: Estimation of the population mean and median using truncation-based ranked set samples. Accepted in J Stat Comput Simul 2012. 10.1080/00949655.2012.662684Google Scholar
- Al-Saleh MF, Al-Kadiri MA: Double ranked set sampling. Stat probability lett 2000,48(2):205–212. 10.1016/S0167-7152(99)00206-0View ArticleGoogle Scholar
- Al-Saleh MF, Al-Omari AI: Multistage ranked set sampling. J Stat Planning and Inference 2002,102(2):273–286. 10.1016/S0378-3758(01)00086-6View ArticleGoogle Scholar
- Ebrahimi N, Pflughoeft K, Soofi E: Two measures of sample entropy. Stat Probability Lett 1994, 20: 225–234. 10.1016/0167-7152(94)90046-9View ArticleGoogle Scholar
- Folks JL, Chhikara RS: The inverse Gaussian distribution and its statistical application-a review. J R Soc, Series B 1998, 40: 263–289.Google Scholar
- Haq A, Shabbir J: A family of ratio estimators for population mean in extreme ranked set sampling using two auxiliary variables. SORT 2010,34(1):45–64.Google Scholar
- Mahdizaheh M, Arghami NR: Efficiency of ranked set sampling in entropy estimation and goodness-of-fit testing for the inverse Gaussian law. J Stat Comput Simul 2010,80(7):761–774. 10.1080/00949650902773551View ArticleGoogle Scholar
- McIntyre GA: A method for unbiased selective sampling using ranked sets. Australian J Agricultural Res 1952, 3: 385–390. 10.1071/AR9520385View ArticleGoogle Scholar
- Mudholkar GS, Tian L: An entropy characterization of the inverse Gaussian distribution and related goodness-of-fit test. J Stat Planning and Inference 2002, 102: 211–221. 10.1016/S0378-3758(01)00099-4View ArticleGoogle Scholar
- Park S, Park D: Correcting moments for goodness of fit tests based on two entropy estimates. J Stat Comput Simul 2003,73(9):685–694. 10.1080/0094965031000070367View ArticleGoogle Scholar
- Seshadri V: The inverse Gaussian distribution: Statistical theory and applications. Springer, New York; 1999.View ArticleGoogle Scholar
- Shannon CE: A mathematical theory of communications. Bell System Technical J 1948,27(379–423):623–656.View ArticleGoogle Scholar
- Takahasi K, Wakimoto K: On the unbiased estimates of the population mean based on the sample stratified by means of ordering. Annals of the Institute of Statistical Mathematics 1968, 20: 1–31. 10.1007/BF02911622View ArticleGoogle Scholar
- Van Es B: Estimating functionals related to a density by class of statistics based on spacing's. Scand J Stat 1992, 19: 61–72.Google Scholar
- Vasicek O: A test for normality based on sample entropy. J Royal Stat Soc B 1976, 38: 54–59.Google Scholar
Copyright
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.