Professional Documents
Culture Documents
discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/281165575
CITATION READS
1 56
1 author:
Suchandan Kayal
National Institute of Technology Rourkela
15 PUBLICATIONS 15 CITATIONS
SEE PROFILE
Some of the authors of this publication are also working on these related projects:
All content following this page was uploaded by Suchandan Kayal on 12 April 2016.
Research Article
Generalized Residual Entropy and Upper Record Values
Suchandan Kayal
Department of Mathematics, National Institute of Technology Rourkela, Rourkela 769008, India
Copyright 2015 Suchandan Kayal. This is an open access article distributed under the Creative Commons Attribution License,
which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
In this communication, we deal with a generalized residual entropy of record values and weighted distributions. Some results on
monotone behaviour of generalized residual entropy in record values are obtained. Upper and lower bounds are presented. Further,
based on this measure, we study some comparison results between a random variable and its weighted version. Finally, we describe
some estimation techniques to estimate the generalized residual entropy of a lifetime distribution.
density function and the survival function of the th upper expressed in terms of that of . Let be a random variable
record value are given by having truncated gamma distribution with density function
1
1 () ( | , ) = ,
() = () , (3) (, ) (7)
( 1)!
1 > 0, > 0, > 0.
() (; ())
() = () = , (4)
=0 ! () For convenience, we denote (, ).
value from (0, 1) distribution. Upper and lower bounds are (9)
obtained. Monotone behaviour of (2) based on the upper 1
record values is investigated. In Section 3, based on (2), = , ( ; ()) + ( )
we study comparisons between a random variable and its
weighted version. We describe some estimation techniques to ln [ (1 (1 (1 )))] ,
estimate the generalized residual entropy of a life distribution
in Section 4. Some concluding remarks have been added since , ( ; ()) = ((1)+1, ln()). This completes
in Section 5. Throughout the paper, we assume that the the proof.
random variables are nonnegative. The terms increasing
and decreasing stand for nondecreasing and nonincreasing, As a consequence of Theorem 2, we get the following
respectively. remark.
where (( 1) + 1, 1).
, ( ; )
In Table 1, we obtain expressions of , ( ; ) for
( ( 1) + 1, ln (1 )) (6)
1 Weibull and Pareto distributions. It is easy to show that
=( ) ln .
(; ln (1 )) ()/ () and ()/() are increasing functions in
0 (see Li and Zhang [18]). Therefore, we have the following
In the following theorem, we show that the generalized theorem whose proof follows along the lines similar to those
residual entropy of the upper record value can be in Theorem 8 of Kayal [11].
Journal of Probability and Statistics 3
Table 1: Expressions of , ( ; ).
Table 2: Bounds of , ( ; ).
PDF Bounds
1 (1)
() + ( ) ln [ ]
( | , ) = , > 0, > 0 ( + 1) ( 1) (+1)(1)
+1 1
() , ( ; ) + ( ) ln ( ), for + > (<)2
1
1 ( + 1)
1 () + ln ( exp ( ))
( | ) = exp ( ), > 0, > 0 +1
1
() , ( ; ) ln(), for + > (<)2
Theorem 4. The th upper record value is increasing generalized residual entropy of upper record values in terms
(decreasing) generalized residual entropy (IGRE (DGRE)) if of .
is IGRE (DGRE).
Definition 7. Let and be two nonnegative random
Note that for th and ( + 1)th upper record values, variables with survival functions () and (), respectively.
()/ () is increasing in 0 (see Kochar [19]) Then, is said to be smaller than in the usual stochastic
+1
st
and hence +1 . Therefore, the following corollary ordering, denoted by , if () (), for all 0.
immediately follows from Theorem 4.
Theorem 8. Let 1 , 2 , . . . be a sequence of upper record
Corollary 5. The ( + 1)th upper record value +1 is IGRE values from a distribution with cumulative distribution func-
(DGRE) if th upper record value is IGRE (DGRE). tion () and probability density function (). Also let ()
be an increasing function. Then, , ( ; ) is increasing
The following theorem provides bounds for the general- (decreasing) in for + > (<)2.
ized residual entropy of the th upper record . We omit
the proof as it follows in an analogous approach similar to Proof. From (8), we have
that of Theorem 11 given in Kayal [11].
, (+1 ; ) , ( ; )
Theorem 6. (a) Let , ( ; ) be finite. If = max{(
1), ln ()} is the mode of ln() (( 1) + 1, 1), then = , (+1
; ()) , ( ; ())
(13)
1 1 +1
, ( ; ) () 1 ( ( (1 )))
+( ) ln 1 1
.
(11) ( ( (1 )))
1
+( ) ln () 1 () ,
Moreover, for + > (<)2,
where () = (/( )) ln (; ln ()) + (( 1)/(
)) ln (1/( )) . ln() (1)
(b) Let = () < , where is the mode of the (14)
distribution with density function (). Also denote the ln() (1)
th upper record value from a sequence of observation from
(0, 1). Then, for + > (<)2, is an increasing (decreasing) function in . Therefore, for
st st
1 () = and V () = , we have ()1 for
, ( ; ) () , ( ; ) + ( ) ln . (12) + > (<)2. Hence, from Theorem 12 of Kayal [11], it can be
proved that , ( ; ()) is increasing (decreasing) for +
As an application of Theorem 6, we obtain bounds of Pareto > (<)2. Now, along the lines of the proof of Theorem 3.7 of
and exponential distributions presented in Table 2. The Zarezadeh and Asadi [20], the proof follows. This completes
following theorem gives the monotone behaviour of the the proof of the theorem.
4 Journal of Probability and Statistics
3. Weighted Distributions the proof of part (a) follows from Theorem 10. Part (b) can
be proved similarly. This completes the proof of the theorem.
To overcome the difficulty to model nonrandomized data set
in environmental and ecological studies, Rao [21] introduced
the concept of weighted distributions. Let the probability Let be a random variable with density function ()
density function of be (), and let () be the nonneg- and cumulative distribution function (). Also let =
ative function with = (()) < . Also let () and () > 0 be finite. Denote the length biased version of
(), respectively, be the probability density function and by . Then, the probability density function of is given
survival function of the weighted random variable , which by
are given by
()
() = . (18)
() ()
() = , (15)
The random variable arises in the study of lifetime analysis
( () | ) () and various probability proportional-to-size sampling prop-
() = . (16) erties. Associated with a random variable , one can define
another random variable with density function
The hazard rate of is
()
() = . (19)
()
() = () . (17)
( () | )
This distribution is known as equilibrium distribution of .
For some results and applications on weighted distributions, The random variables and are weighted versions of
one may refer to Di Crescenzo and Longobardi [22], Gupta with weight function () = and () = 1/ (),
and Kirmani [23], Kayal [9], Maya and Sunoj [24], Navarro et respectively. The following corollary is a consequence of
al. [25], and Patil [26]. In the present section, we obtain some Theorem 12.
comparison results based on the generalized residual entropy
between a random variable and its weighted version. We need Corollary 13. Let be DFR. Then, for all 0,
the following definition in this direction. (a) , (; ) (), ( ; ) for + > (<)2;
Definition 9. A random variable with hazard rate () (b) , (; ) (), ( ; ) for + > (<)2.
is said to have a decreasing (increasing) failure rate (DFR
(IFR)), if () is decreasing (increasing) in 0. 4. Estimation
Theorem 10. Let and be two random variables with In this section, we discuss the problem of estimation of
cumulative distribution functions () and (), probability the generalized residual entropy of a statistical distribution
density functions () and (), survival functions () based on upper record values. Here, we consider exponential
and (), and hazard rates () and (), respectively. If distribution. It has various applications in practice. Let
() (), for all 0, and either () or () is DFR, follow exponential distribution with mean . Then, from (2),
then , (; ) (), (; ), for + > (<)2. we obtain
Proof. It is not difficult to see that () (), for all 0, Also the uniformly minimum variance unbiased estimator
when either (() | ) or () is decreasing. Now, (umvue) of , (; ) can be obtained as
Journal of Probability and Statistics 5
mv = ( + 1 ) ln (
)
[8] S. Kayal, On generalized dynamic survival and failure entropies
, of order (, ), Statistics and Probability Letters, vol. 96, pp. 123
exp ( ())
(22) 132, 2015.
1 [9] S. Kayal, Some results on dynamic discrimination measures of
( ) ln ( + 1) ,
order (, ), Hacettepe Journal of Mathematics and Statistics,
vol. 44, pp. 179188, 2015.
where () is a digamma function. To illustrate the esti- [10] S. Kayal, Characterization based on generalized entropy of
mation techniques developed in this section, we consider order statistics, Communications in StatisticsTheory and
simulated data from exponential distribution with mean 1. Methods. In press.
In this purpose, we use Monte-Carlo simulation. [11] S. Kayal, Some results on a generalized residual entropy based
on order statistics, Statistica. In press.
Example 14. In this example, we consider a simulated sample [12] S. Kayal and P. Vellaisamy, Generalized entropy properties of
of size = 5 from the exponential distribution with mean 0.5. records, Journal of Analysis, vol. 19, pp. 2540, 2011.
The simulated upper records are as follows: [13] V. Kumar and H. C. Taneja, Some characterization results on
generalized cumulative residual entropy measure, Statistics &
0.265410, 0.637725, 0.688878, 0.791721, Probability Letters, vol. 81, no. 8, pp. 10721077, 2011.
(23) [14] M. M. Sati and N. Gupta, On partial monotonic behaviour of
2.114831. Varma entropy and its application in coding theory, Journal of
the Indian Statistical Association. In press.
ml = 6.64759
Based on these upper record values, we have , [15] N. Glick, Breaking records and breaking boards, The American
mv = 6.06166 when = 1.2 and = 1.5.
and Mathematical Monthly, vol. 85, no. 1, pp. 226, 1978.
,
[16] J. Ahmadi and N. R. Arghami, Comparing the Fisher informa-
tion in record values and iid observations, Statistics, vol. 37, no.
5. Concluding Remarks 5, pp. 435441, 2003.
In this paper, we consider generalized residual entropy due [17] B. C. Arnold, N. Balakrishnan, and H. N. Nagaraja, Records,
to Varma [5] of record values and weighted distributions. We John Wiley & Sons, New York, NY, USA, 1998.
obtain some results on monotone behaviour of this measure [18] X. Li and S. Zhang, Some new results on Renyi entropy of
in upper record values. Some bounds are obtained. Further, residual life and inactivity time, Probability in the Engineering
and Informational Sciences, vol. 25, no. 2, pp. 237250, 2011.
some comparison results between a random variable and its
weighted version based on the generalized residual entropy [19] S. C. Kochar, Some partial ordering results on record values,
Communications in StatisticsTheory and Methods, vol. 19, no.
are studied. Finally, two estimators of the generalized residual
1, pp. 299306, 1990.
entropy of exponential distribution have been described.
[20] S. Zarezadeh and M. Asadi, Results on residual Renyi entropy
of order statistics and record values, Information Sciences, vol.
Conflict of Interests 180, no. 21, pp. 41954206, 2010.
[21] C. R. Rao, On discrete distributions arising out of methods of
The author declares that there is no conflict of interests ascertainment, Sankhya, vol. 27, pp. 311324, 1965.
regarding the publication of this paper. [22] A. Di Crescenzo and M. Longobardi, On weighted residual and
past entropies, Scientiae Mathematicae Japonicae, vol. 64, no. 2,
References pp. 255266, 2006.
[23] R. C. Gupta and S. N. U. A. Kirmani, The role of weighted
[1] C. E. Shannon, A mathematical theory of communication, The distributions in stochastic modeling, Communications in
Bell System Technical Journal, vol. 27, no. 3, pp. 379423, 1948. StatisticsTheory and Methods, vol. 19, no. 9, pp. 31473162,
[2] J. N. Kapur, Generalized entropy of order alpha and type beta, 1990.
in Proceedings of the Mathematical Seminar, pp. 7894, New [24] S. S. Maya and S. M. Sunoj, Some dynamic generalized
Delhi, India, 1967. information measures in the context of weighted models,
[3] A. Renyi, On measures of entropy and information, in Statistica, vol. 68, no. 1, pp. 7184 (2009), 2008.
Proceedings of the 4th Berkeley Symposium on Mathematics, [25] J. Navarro, Y. del Aguila, and J. M. Ruiz, Characterizations
Statistics and Probability 1960, vol. 1, pp. 547461, University of through reliability measures from weighted distributions, Sta-
California Press, Berkeley, Calif, USA, 1961. tistical Papers, vol. 42, no. 3, pp. 395402, 2001.
[4] C. Tsallis, Possible generalization of Boltzmann-Gibbs statis- [26] G. P. Patil, Weighted distributions, in Encyclopedia of Envi-
tics, Journal of Statistical Physics, vol. 52, no. 1-2, pp. 479487, ronmetrics, A. H. El-Shaarawi and W. W. Piegorsch, Eds., John
1988. Wiley & Sons, Chichester, UK, 2002.
[5] R. S. Varma, Generalization of Renyis entropy of order ,
Journal of Mathematical Sciences, vol. 1, pp. 3448, 1966.
[6] N. Ebrahimi, How to measure uncertainty in the residual life
time distribution, Sankhya, vol. 58, no. 1, pp. 4857, 1996.
[7] M. Asadi, N. Ebrahimi, and E. S. Soofi, Dynamic generalized
information measures, Statistics & Probability Letters, vol. 71,
no. 1, pp. 8598, 2005.
Advances in Advances in Journal of Journal of
Operations Research
Hindawi Publishing Corporation
Decision Sciences
Hindawi Publishing Corporation
Applied Mathematics
Hindawi Publishing Corporation
Algebra
Hindawi Publishing Corporation
Probability and Statistics
Hindawi Publishing Corporation
http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014 http://www.hindawi.com Volume 2014
International
Journal of Journal of
Mathematics and
Mathematical
Discrete Mathematics
Sciences