scispace - formally typeset

Journal ArticleDOI

Constrained Epsilon-Minimax Test for Simultaneous Detection and Classification

01 Dec 2011-IEEE Transactions on Information Theory (IEEE)-Vol. 57, Iss: 12, pp 8055-8071

TL;DR: A constrained epsilon-minimax test is proposed to detect and classify nonorthogonal vectors in Gaussian noise, with a general covariance matrix, and in presence of linear interferences to minimizes the maximum classification error probability subject to a constraint on the false alarm probability.

AbstractA constrained epsilon-minimax test is proposed to detect and classify nonorthogonal vectors in Gaussian noise, with a general covariance matrix, and in presence of linear interferences. This test is epsilon-minimax in the sense that it has a small loss of optimality with respect to the purely theoretical and incalculable constrained minimax test which minimizes the maximum classification error probability subject to a constraint on the false alarm probability. This loss is even more negligible as the signal-to-noise ratio is large. Furthermore, it is also an epsilon-equalizer test since its classification error probabilities are equalized up to a negligible difference. When the signal-to-noise ratio is sufficiently large, an asymptotically equivalent test with a very simple form is proposed. This equivalent test coincides with the generalized likelihood ratio test when the vectors to classify are strongly separated in term of Euclidean distance. Numerical experiments on active user identification in a multiuser system confirm the theoretical findings.

Topics: Likelihood-ratio test (58%), Minimax (54%), Score test (53%), Gaussian noise (53%), Noise measurement (52%)

Summary (5 min read)

Introduction

  • This problem has many applications including radar and sonar signal processing [1], image processing [2], speech segmentation [3], [4], integrity monitoring of navigation systems [5], quantitative nondestructive testing [6], network monitoring [7] and digital communication [8] among others.
  • The unknown nuisance vector belongs to the nuisance parameter subspace spanned by the columns of a known matrix .

A. Relation to Previous Work

  • From the statistical point of view, this problem of simultaneous detection/classification can be viewed as an hypotheses testing problem between several composite hypotheses [9], [10].
  • The first approach to the design of statistical detection and classification tests is the uncoupled design strategy where detection performance is optimized under the false alarm constraint and the classification is gated by this optimal detection.
  • All the above mentioned probabilities generally vary as a function of both the vector and the nuisance parameter .
  • This strategy is optimal only in some cases.
  • These results are not yet extended to the case of multiple hypotheses testing with a constraint on the false alarm probability.

B. Motivation of the Study

  • The design of the optimal constrained minimax test mainly depends on three major points: 1) the geometric complexity of the vector constellation, 2) the covariance matrix of the Gaussian noise and 3) the presence of nuisance parameters.
  • In the simplest case, the vectors are orthogonal and have the same norm (the least complex vector constellation), the covariance matrix is the identity matrix (possibly multiplied by a known scalar) and there is no nuisance parameter .
  • This solution depends on some unknown coefficients, namely the optimal weights and the threshold.
  • The common solution, namely the parity space approach, involves two steps.
  • When the optimal statistical test is unknown or intractable, it is often assumed that the optimal weights are equal (since it is the least informative a priori choice) and the threshold is tuned to satisfy the false alarm constraint.

C. Contribution and Organization of the Paper

  • The first contribution is the design of a constrained -minimax detection/classification test solving the detection/classification problem in the case of nonorthogonal vectors with linear nuisance parameters and an additive Gaussian noise with a known general covariance matrix.
  • It is also shown that this test coincides with a constrained -equalizer Bayesian test which equalizes the classification error probabilities over the alternative hypotheses up to a constant .
  • This map is also used to calculate in advance the asymptotic maximum classification error probability of the constrained -minimax test as the Signal-to-Noise Ratio (SNR) tends to infinity.
  • Finally, it is shown that the MGLRT is -optimal when the mutual geometry between the hypotheses is very simple, i.e., when each vector has at most one other vector nearest to it in term of Euclidean distance.
  • In general, the MGLRT is suboptimal and the loss of optimality may be significant.

II. PROBLEM STATEMENT

  • This section presents the multiple hypotheses testing problem which consists in detecting and classifying a vector in the presence of linear nuisance parameters.
  • A new optimality criterion, namely the constrained -minimax criterion, is introduced and motivated.

A. Multiple Hypotheses Testing

  • Without any loss of generality, it is assumed that the noise vector follows a zero-mean Gaussian distribution .
  • In fact, it is always possible to multiply (1) on the left by the inverse square-root matrix of to obtain the linear Gaussian model with the vector , the nuisance matrix and a Gaussian noise having the covariance matrix .
  • The following condition of separability is assumed to be satisfied: (3) In other words, it is assumed that the intersection of the two linear manifolds and (which are parallel to each other) is an empty set for all (two parallel linear manifolds with nonempty intersection are equal).
  • A test function for the multiple hypotheses is a -dimensional vector function defined on such that and Given , the test function decides the hypothesis if and only if, also known as Definition 1.
  • The maximum classification error probability for the test function is denoted.

B. Constrained Epsilon-Minimax Test

  • As mentioned in the introduction, the constrained minimax criterion given in [18] is a very natural criterion for problem (2).
  • Hence, to overcome this difficulty, it makes sense to consider constrained -minimax tests, i.e., tests that approximate optimal minimax test with a small loss, say , of optimality.
  • There exists a positive function satisfying as such that for any other test function .
  • Obviously, Definition 2 assumes that the positive constant is (very) small.
  • In some cases, it is possible to get (see Section VI) but, generally, because of the vector constellation complexity.

III. EPSILON-MINIMAX TEST FOR COMPOSITE HYPOTHESES

  • This section introduces the constrained -equalizer Bayesian test of level .
  • The first step to design the constrained -equalizer Bayesian test between composite hypotheses in presence of nuisance parameters consists in eliminating these unknown parameters.
  • Proposition 2 shows that this elimination, based on the nuisance parameters rejection, leads to a reduced decision problem between simple statistical hypotheses.

A. Constrained Epsilon-Equalizer Test

  • Let us recall the definition of the constrained Bayesian test before introducing the definition of the constrained -equalizer test.
  • Let be a probability distribution over called the a priori distribution.
  • The constrained Bayesian test function of level associated to is given by (6) and for (7) where the threshold is selected to satisfy the constraint (8) The following -equalization criterion serves to design a constrained -minimax test.
  • Definition 3: A test function is a constrained -equalizer test between the hypotheses in the class if the following condi- tions are fulfilled : i) ; ii).

B. Reduction to Epsilon-Equalizer Test for Simple Hypotheses

  • The presence of linear nuisance parameters complicates the statistical decision problem.
  • The family of distributions for and remains invariant (see [10] for details and definitions) under the group of translations which induces in the parameter space the group that preserves all the sets , i.e., for all and .
  • The following proposition shows that a constrained -equalizer Bayesian test for the reduced problem (13) is a constrained -minimax test for the initial problem (2).
  • Define the -dimensional unit simplex Proof: Let for all .
  • This comes from the fact that the rejection principle is equivalent to assume that the parameter follows under a degenerate a priori distribution whose support consists of only one value for all .

A. Principle of Epsilon-Equalization

  • Designing a constrained -equalizer Bayesian test for the problem (13) is difficult for mainly three reasons: i) the common value of the classification error probabilities is unknown, ii) the weight vector ensuring the -equalization of the classification error probabilities is not easily calculable and iii) the threshold must be chosen accordingly to the prescribed false alarm probability .
  • Loosely speaking, the separability map is a graph [34], [35] whose topology is deduced from the Euclidean distances between the vectors to be classified.
  • The neighborhood of characterizes the maximum classification error probability of .
  • Thus, it is proposed to solve an original linear programming problem whose solution gives the remaining weights.
  • To the authors’s knowledge, such an ap- proach has never been addressed before in the literature.

B. Mutual Geometry for Simple Gaussian Hypotheses

  • The distance between the null hypothesis and the alterna- tive one is defined as (15) where is a real depending on the prescribed false alarm probability .
  • Suppose that the component contains elements with, also known as Lemma 1.
  • Then, according to the definition of the critical value, and one obtain .
  • When the component is not a star graph, the least separable vector is not easily identifiable and it becomes necessary to study in details the internal geometry of the component in order to calculate .
  • The separability map and its two components and are represented in Fig.

C. Constrained Epsilon-Equalizer Test for Simple Hypotheses

  • To derive the constrained -equalizer test, some bounds on the classification error probabilities and also on the false alarm probability are used.
  • Finally, constraint (28) means that the weight of the noncritical vector must not exceed a certain level imposed by the prescribed false alarm probability.
  • Under the assumptions A1) and A2), there exist a weight vector given by (23) and (29) and a threshold (30) where , for which the test is a constrained -equalizer test between such that, also known as Theorem 1.
  • The optimal weight vector is computed by using an iterative algorithm.

V. ASYMPTOTICALLY EQUIVALENT TEST

  • This section proposes some asymptotically equivalent tests to the constrained -minimax one given in Theorem 1.
  • These tests have the same asymptotic maximum classification error but their form is simpler.

A. Simplified Asymptotic Form of the Epsilon-Minimax Test

  • Let be the positive real where with and Let be cut into the three disjoint subsets and such that where denotes the component containing .
  • Let be defined by for all where and let be the threshold (34) Proposition 4: Under the assumptions A1) and A2), the exact comparison between and is not relevant.
  • It must be noted that is not necessarily a constrained -equalizer test.

B. Optimality of the MGLRT

  • Let be the uniform weight vector such that for all .
  • Let Define as the number of vectors at distance from and Proposition 5: The test , with the threshold satisfies and (35) Proof: From (55) applied to the weight vector with the threshold , the authors get for and for .
  • Generally, the test is not a constrained -minimax test because .
  • Corollary 2: Under the assumptions A1), A2) and A3), the test such that is asymptotically equivalent to the constrained -minimax test as (36) Proof: Under the assumptions A1) and A3), the authors get .
  • It is important to note that the assumption A3) is very severe in practice since it imposes very strict requirements on the mutual geometry between the hypotheses.

A. Detection of a New User Entrance

  • In order to make the simulation free of secondary details, let us consider the plainest, yet general enough, model of Direct-Sequence/Code-Division Multiple-Access (DS/CDMA) involving real signatures and Binary Phase Shift Keying (BPSK) data transmission [29], [30].
  • Following [31], the fact that is not explicitly taken into account.
  • Strictly speaking, the goal is to estimate both the entry time, the signature waveform vector and the transmitted bit of the new user.
  • This is a change detection/classification problem between hypotheses [31].
  • In contrast to the sequential strategy, the repeated Fixed Size Sample (FSS) one is easily applicable to systems with a variable structure for quickest detection and classification of changes.

B. Simulation Results

  • For simplicity, it is assumed that the signatures are chosen such that the rejection mechanism (see Subsection III.B) leads to the reduced model where and is either or one of the following vectors: for .
  • It has the same separability values and critical values than .
  • This zone contains the maximum classification error probability .
  • The false alarm zone of the constrained -minimax test is the region defined by the two following curves.
  • The second one, at the bottom of the region, is the lower bound given in (58).

VII. CONCLUSION

  • This paper has proposed a constrained -minimax test to detect and classify nonorthogonal vectors in presence of linear nuisance parameters.
  • Proposition 2 shows that these nuisance parameters can be rejected without any significant loss of optimality.
  • Theorem 1 proposes a test which classifies the nonorthogonal vectors obtained after this rejection by equalizing the classification error probabilities up to a small constant under a constraint on the false alarm probability.
  • The test design is based on some weighting coefficients which are computed by solving a linear programming problem deduced from the separability map.
  • Finally, Proposition 5 proves that the proposed test clearly outperforms the famous MGLRT when the vector constellation is too complex.

Did you find this useful? Give us your feedback

...read more

Content maybe subject to copyright    Report

HAL Id: hal-00749180
https://hal.archives-ouvertes.fr/hal-00749180
Submitted on 6 Nov 2012
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Constrained Epsilon-Minimax Test for Simultaneous
Detection and Classication
Lionel Fillatre
To cite this version:
Lionel Fillatre. Constrained Epsilon-Minimax Test for Simultaneous Detection and Classication.
IEEE Transactions on Information Theory, Institute of Electrical and Electronics Engineers, 2011, 57
(12), pp.8055-8071. �10.1109/TIT.2011.2170114�. �hal-00749180�

IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011 8055
Constrained Epsilon-Minimax Test for Simultaneous
Detection and Classification
Lionel Fillatre
Abstract—A constrained epsilon-minimax test is proposed to de-
tect and classify nonorthogonal vectors in Gaussian noise, with a
general covariance matrix, and in presence of linear interferences.
This test is epsilon-minimax in the sense that it has a small loss of
optimality with respect to the purely theoretical and incalculable
constrained minimax test which minimizes the maximum classifi-
cation error probability subject to a constraint on the false alarm
probability. This loss is even more negligible as the signal-to-noise
ratio is large. Furthermore, it is also an epsilon-equalizer test since
its classification error probabilities are equalized up to a negligible
difference. When the signal-to-noise ratio is sufficiently large, an
asymptotically equivalent test with a very simple form is proposed.
This equivalent test coincides with the generalized likelihood ratio
test when the vectors to classify are strongly separated in term of
Euclidean distance. Numerical experiments on active user identifi-
cation in a multiuser system confirm the theoretical findings.
Index Terms—Constrained minimax test, generalized likelihood
ratio test, linear nuisance parameters, multiple hypothesis testing,
statistical classification, user activity detection.
I. INTRODUCTION
T
HE problem of detecting and classifying a vector in noisy
measurements under uncertainty of vector presence often
appears in engineering applications. This problem has many
applications including radar and sonar signal processing [1],
image processing [2], speech segmentation [3], [4], integrity
monitoring of navigation systems [5], quantitative nondestruc-
tive testing [6], network monitoring [7] and digital communi-
cation [8] among others. This paper deals with the following
detection and classification problem. It is assumed that a mea-
surement vector
consists of either a vector of interest
(for example, a target or an anomaly) plus an unknown nuisance
vector in additive Gaussian noise, or just an unknown nuisance
vector in additive Gaussian noise. If present, the vector of in-
terest must be detected and classified. The unknown nuisance
vector belongs to the nuisance parameter subspace spanned by
the columns of a known
matrix . Hence, the observa-
tion model has the form
(1)
Manuscript received November 29, 2009; revised June 08, 2011; accepted
June 28, 2011. Date of current version December 07, 2011. This work was sup-
ported in part by the French National Agency of Research under Grant ANR-08-
SECU-013-02.
The author is with the ICD, LM2S, Université de Technologie de Troyes
(UTT), UMR STMR, CNRS 6279, BP 2060, 10010, Troyes, France (e-mail:
lionel.fillatre@utt.fr).
Communicated by M. Lops, Associate Editor for Detection and Estimation.
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org.
Digital Object Identifier 10.1109/TIT.2011.2170114
where both the nuisance parameter vector
and the
vector
are unknown and deterministic. The zero-mean
Gaussian noise vector
has the known positive definite general
covariance matrix
. The vector belongs to the known set of
different vectors , also called
the vector constellation. The relation between the dimension
of the observed vector and the number of non-null vectors
is arbitrary but it is assumed that
.
Three objectives are aimed to be achieved: i) vector detection
which is to decide if
for a given false alarm probability
(probability to declare an alarm when the observation vector is
anomaly-safe), ii) vector classification which is to specify the
actual index
of the vector and iii) insensitivity to the nui-
sance parameters which consists of taking the simultaneous de-
tection/classification decision independently from the value of
the unknown vector
.
A. Relation to Previous Work
From the statistical point of view, this problem of simulta-
neous detection/classification can be viewed as an hypotheses
testing problem between several composite hypotheses [9], [10].
The goal is to design a statistical test which achieves the above
mentioned three objectives according to a prefixed criterion of
optimality.
The first approach to the design of statistical detection and
classification tests is the uncoupled design strategy where detec-
tion performance is optimized under the false alarm constraint
and the classification is gated by this optimal detection. On the
one hand, the classical Neyman-Pearson criterion of vector de-
tection [10] states that it is desirable to minimize the probability
to miss the target subject to a constraint on the false alarm prob-
ability. On the other hand, in terms of target classification, it
is desirable to minimize the probabilities to badly classify the
target. All the above mentioned probabilities generally vary as
a function of both the vector
and the nuisance parameter .
Hence, the uniform minimization of these probabilities with re-
spect to
and is in general impossible. There is no guarantee
that the global performance of this uncoupled strategy will be
acceptable. Consequently, a different approach must be taken,
namely the coupled design strategies for detection and classifi-
cation. These strategies have been studied by only a few authors.
Pioneering works include the papers [11]–[14]. The common
ground in each of these studies is the Bayesian point of view,
i.e., prior probabilities are assigned to all the parameters so that
average performance can be optimized. The problem of simulta-
neous detection and classification using a combination of a gen-
eralized likelihood ratio test and a maximum-likelihood classi-
fier is studied in [15], [16]. This strategy is optimal only in some
cases.
0018-9448/$26.00 © 2011 IEEE

8056 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011
Contrary to a purely Bayesian criterion which needs a com-
plete statistical description of the problem, the minimax crite-
rion is well adapted to detection problems where some param-
eters are deterministic and unknown (typically the nuisance pa-
rameter
) and the appearance probability of each vector is
unknown. This criterion consists of minimizing the largest prob-
ability to make a decision error (typically a miss detection or a
classification error). In [17], the generalized-likelihood ratio test
approach is extended to multiple composite hypothesis testing,
by breaking the problem into a sequence of binary composite
hypothesis tests. In some cases, sufficient conditions for min-
imax optimality of this strategy are provided. Lastly, a general
framework to design minimax tests with a prefixed level of false
alarm, namely the constrained minimax tests, between multiple
hypotheses composed of a finite number of parameters is estab-
lished in [18].
It must be mentioned that some interesting papers like [19],
[20] study the asymptotic performance of Bayesian tests be-
tween multiple hypotheses in absence of constraints on the false
alarm probability. When the number of observations is very
large, these papers show that the error probabilities depend only
on the Kullback-Leibler information between the hypotheses.
These results are not yet extended to the case of multiple hy-
potheses testing with a constraint on the false alarm probability.
B. Motivation of the Study
The design of the optimal constrained minimax test mainly
depends on three major points: 1) the geometric complexity
of the vector constellation, 2) the covariance matrix of the
Gaussian noise and 3) the presence of nuisance parameters. To
underline the importance of these points, the three following
cases must be distinguished.
In the simplest case, the vectors
are orthogonal and have
the same norm (the least complex vector constellation), the co-
variance matrix
is the identity matrix (possibly multiplied by
a known scalar) and there is no nuisance parameter
. The op-
timal solution of the problem is given in [9], [21]: this is the
so-called
-slippage problem.
In a more difficult case, the vectors
are not orthogonal
and/or they have different norms (the most complex vector con-
stellation), the Gaussian noise has a general covariance matrix
(not necessarily diagonal) and there is no nuisance parameter
. The theoretical optimal solution is given by the constrained
minimax test [18] but it is generally intractable. This optimal
test compares the maximum of weighted likelihood ratios to
a threshold to take its decision. Although the existence of the
optimal solution is established, this solution depends on some
unknown coefficients, namely the optimal weights and the
threshold. Furthermore, the presence of a general covariance
matrix
plays a significant role in the calculation of the
optimal weights, even if the vector constellation is simple.
In fact, it is always possible to get a diagonal covariance
matrix after prefiltering but this operation may involve that
the vector constellation becomes more complex. For example,
orthogonal vectors of interest may be no longer orthogonal
after prefiltering. For all these reasons, it is often impossible
to easily calculate the optimal weights and, even, to reduce
the number of weights to be determined by using invariance
principles [10], [22]. The optimization problem to be solved for
calculating the optimal weights is highly nonlinear and leads to
a combinatorial explosion.
Example 1 (Slippage Problem With Unstructured Noise):
This example is directly related to the important problem of
detecting outliers in multivariate normal data [23]. Let
be a
Gaussian random vector with zero mean and a known general
covariance matrix under hypothesis . Under hypothesis
, the th component of has the known mean .
Hence, the vector
to detect and classify corresponds to
where denotes the vector with a 1 in the th coordinate and
0’s elsewhere. Since the covariance is known but it differs from
the identity matrix, it is no longer possible to use the famous
principle of invariance to solve such a slippage problem [9].
The optimal solution is not known up to now. Example 6 shows
that the results proposed in this paper can be used to solve this
slippage problem.
In the most difficult case, the vectors ’s are not orthog-
onal and/or they have different norms, the Gaussian noise has
a general covariance matrix
and there is an unknown nui-
sance parameter
. To our knowledge, the optimal constrained
minimax test is unknown in this case. The main reasons which
explain this lack of results are the followings. First, the presence
of linear nuisance parameters certainly complicates the mutual
geometry between the vectors. Next, there is an unavoidable an-
tagonism between the detection and classification performances
of the test. For example, to get small classification errors, it is
necessary to accept a loss of sensibility for the probability of de-
tection. The tradeoff between these two requirements is essen-
tially based on the worst case of detection and the worst case of
classification which are generally difficult to identify. Finally, as
underlined in [24]–[27], the analytic calculation of the miss de-
tection probability and the classification error probabilities are
intractable, which makes difficult the derivation of an optimal
test.
Example 2 (Integrity Monitoring of Navigation Systems): Let
be a Gaussian random vector with the known covariance ma-
trix . Under hypothesis , its mean is where is the user
unknown parameters and
is a matrix describing the measure-
ment system [5]. Under hypothesis
is contaminated by
a scalar error with intensity
. The common solution, namely
the parity space approach, involves two steps. First, the user un-
known parameters are eliminated by projecting
on the null-
space of
. This null-space is called the parity space in the ana-
lytical redundancy literature [28]. Next, the error is detected and
classified (isolated) directly in the parity space. Unfortunately,
the first step may generate some linear dependencies between
the possible error signatures in the parity space. In this case, the
problem is not theoretically solved. Example 7 shows that this
paper proposes a solution to this problem.
Example 3 (New User Identification in a Multiuser System):
In a multiuser system, after chip-matched filtering and chip rate
sampling, the received signal vector under hypothesis
is
modeled as
where the th column of the ma-
trix is the normalized unit energy signature waveform vector

FILLATRE: CONSTRAINED EPSILON-MINIMAX TEST FOR SIMULTANEOUS DETECTION AND CLASSIFICATION 8057
of user
is the diagonal matrix of user amplitudes and
is the vector whose th component is the antipodal symbol,
or 1, transmitted by user [29], [30]. The random vector
has zero mean and the known covariance matrix where
is the identity matrix of size . Under hypothesis , the
vector
is added to , i.e., a new user with the sig-
nature
emits the symbol with the amplitude .Itis
assumed that belongs to a finite set of predefined nonorthog-
onal signatures. The multiple-access interferences
can be
eliminated by using the above mentioned parity space approach
[31]. The goal is to detect the new user arrival and to identify
the waveform
. This is a difficult problem, especially when
the common length
of each user’s signature is shorter than
the total number of simultaneously active users [32]. Section VI
will further elaborate upon this example.
When the optimal statistical test is unknown or intractable,
it is often assumed that the optimal weights are equal (since
it is the least informative
a priori choice) and the threshold is
tuned to satisfy the false alarm constraint. The resulting test is
called the
-ary Generalized Likelihood Ratio Test (MGLRT)
between equally probable hypotheses. It must be noted that the
optimality proof of the MGLRT is still an open problem.
C. Contribution and Organization of the Paper
The first contribution is the design of a constrained
-min-
imax detection/classification test solving the detection/clas-
sification problem in the case of nonorthogonal vectors with
linear nuisance parameters and an additive Gaussian noise with
a known general covariance matrix. This test is based on the
maximum of weighted likelihood ratios, i.e., it is a Bayesian
test associated to some specific weights. It is
-optimal (under
mild assumptions) in the sense that it is optimal with a loss
of a small part, say
, of optimality with respect to the purely
theoretical minimax test. This loss of optimality, which is the-
oretically bounded, is unavoidable since the purely theoretical
minimax test is intractable due to the difficulties above men-
tioned. This loss is even more negligible as the signal-to-noise
ratio is large. It is also shown that this test coincides with
a constrained
-equalizer Bayesian test which equalizes the
classification error probabilities over the alternative hypotheses
up to a constant .
Secondly, an algorithm is proposed to compute the optimal
weights of the proposed test with a reasonable numerical com-
plexity. This algorithm is based on a graph, namely the separa-
bility map, describing the mutual geometry between the vectors.
This separability map serves to identify the least separable vec-
tors, making possible the design of the associated constrained
-minimax test. This map is also used to calculate in advance
the asymptotic maximum classification error probability of the
constrained -minimax test as the Signal-to-Noise Ratio (SNR)
tends to infinity. Moreover, in the case of large SNR, an asymp-
totically equivalent test is proposed whose optimal weights have
a very simple form.
Finally, it is shown that the MGLRT is
-optimal when the
mutual geometry between the hypotheses is very simple, i.e.,
when each vector has at most one other vector nearest to it in
term of Euclidean distance. In general, the MGLRT is subop-
timal and the loss of optimality may be significant.
The paper is organized as follows. Section II starts with the
problem statement and introduces the statistical framework that
will be used in this paper, including the presentation of the con-
strained
-minimax criterion. Section III describes the general
methodology to reduce the detection/classification problem
between multiple hypotheses with nuisance parameters to a
detection/classification problem between multiple hypotheses
without nuisance parameters. This reduction is based on the fact
that it is sufficient to design a constrained
-equalizer Bayesian
test to get the constrained
-minimax one. Section IV defines
the separability map and proposes the main theorem of this
paper which establishes the constrained
-equalizer Bayesian
test. The proof of this theorem is given in Appendix A. The
false alarm probability of this test is calculated in Appendix B.
Section V gives an asymptotically equivalent test to the con-
strained
-minimax as the SNR tends to infinity. The derivation
of this test is given in Appendix C. It is also shown that the
MGLRT is asymptotically
-optimal when the mutual geometry
between the hypotheses is very simple. Section VI deals with a
practical problem, namely the identification of a new active user
in a multiuser system, showing the efficiency of the proposed
-optimal test. Finally, Section VII concludes this paper.
II. PROBLEM STATEMENT
This section presents the multiple hypotheses testing problem
which consists in detecting and classifying a vector in the pres-
ence of linear nuisance parameters. A new optimality criterion,
namely the constrained
-minimax criterion, is introduced and
motivated.
A. Multiple Hypotheses Testing
The observation model has the form (1). Without any loss
of generality, it is assumed that the noise vector
follows a
zero-mean Gaussian distribution
. In fact, it is always
possible to multiply (1) on the left by the inverse square-root
matrix
of to obtain the linear Gaussian model with the
vector , the nuisance matrix and a Gaussian noise
having the covariance matrix
. It is also assumed that is a
full-column rank matrix. If the matrix
does not satisfy this
assumption, it suffices to keep the maximum number of linear
independent columns to get a full-column rank matrix spanning
the same linear space. It is then desirable to solve the multiple
Gaussian hypotheses testing problem between the statistical hy-
potheses
(2)
where
. The following condition of
separability is assumed to be satisfied:
(3)
In other words, it is assumed that the intersection of the two
linear manifolds
and (which are parallel to each other)
is an empty set for all
(two parallel linear manifolds with
nonempty intersection are equal). Here, the parameter set
associated to hypothesis is not a singleton, hence, is
called a composite hypothesis [33]. Otherwise, the hypothesis is
simple and it is identified by the absence of an underscore, say

8058 IEEE TRANSACTIONS ON INFORMATION THEORY, VOL. 57, NO. 12, DECEMBER 2011
. The set of decision strategies for the -ary hypotheses
testing problem (2) is specified by the set of test functions.
Definition 1: A test function
for the mul-
tiple hypotheses
is a -dimensional vector
function defined on such that and
Given , the test function decides the hypothesis
if and only if . In this case, the other components
, are zero. The study of randomized test functions
(a random test function satisfies
for some )
is not considered in this paper since the probability distribution
of
is continuous whatever the true hypothesis. The average
performance of a particular test function
is determined by
the
functions where
stands for the expectation of when follows the distribution
. The false alarm probability function is given by
when . The function for
describes the probability of miss detection. When
is the classification error proba-
bility. The maximum classification error probability for the test
function
is denoted
For a test where is a given superscript, all the above men-
tioned notations are completed by the superscript
. Let be
the set of test functions whose maximum
false alarm probability is less or equal to
B. Constrained Epsilon-Minimax Test
As mentioned in the introduction, the constrained minimax
criterion given in [18] is a very natural criterion for problem
(2). Unfortunately, as underlined in [18], since the geometry of
the set
may be very complex, it is impossible
to infer the structure of the minimax test. Hence, to overcome
this difficulty, it makes sense to consider constrained
-minimax
tests, i.e., tests that approximate optimal minimax test with a
small loss, say
, of optimality.
Definition 2: A test function is a con-
strained
-minimax test in the class between the hypotheses
if the following conditions are fulfilled :
i)
;
ii) There exists a positive function
satisfying
as such that
for any other test function .
Obviously, Definition 2 assumes that the positive constant
is
(very) small. In some cases, it is possible to get
(see
Section VI) but, generally,
because of the vector con-
stellation complexity. Contrary to a purely constrained minimax
test, the design of a constrained
-minimax test tolerates small
errors on the classification error probabilities. Hence, it becomes
possible to use lower and upper bounds on these probabilities in
order to evaluate the statistical performances of the test. As un-
derlined in [25], the exact calculation of these probabilities is
generally intractable.
III. E
PSILON-MINIMAX TEST FOR COMPOSITE HYPOTHESES
This section introduces the constrained
-equalizer Bayesian
test of level . Proposition 1 shows that such a test is neces-
sarily a constrained
-minimax one. The first step to design the
constrained
-equalizer Bayesian test between composite hy-
potheses in presence of nuisance parameters consists in elim-
inating these unknown parameters. Proposition 2 shows that
this elimination, based on the nuisance parameters rejection,
leads to a reduced decision problem between simple statistical
hypotheses.
A. Constrained Epsilon-Equalizer Test
Let us recall the definition of the constrained Bayesian test
before introducing the definition of the constrained
-equalizer
test. Let
be a probability distribution over called the a priori
distribution. For all
, this distribution induces
some a priori distributions on the linear manifolds and
some a priori probabilities
such that
(4)
Let
be the probability density function (pdf) of the ob-
servation vector
following the distribution . To each
hypothesis is associated the weighted pdf (see de-
tails in [10]) defined by
Let be the weighted log-likelihood ratio defined by
(5)
for
. The constrained Bayesian test function
of level associated to is given by
(6)
and for
(7)

Citations
More filters

Journal ArticleDOI
TL;DR: A suboptimal CUSUM-type transient change detection algorithm, based on a subclass of truncated Sequential Probability Ratio Tests, is proposed, and the optimization of the proposed algorithm in this subclass leads to a specially designed Finite Moving Average Test.
Abstract: This paper addresses the detection of a suddenly arriving dynamic profile of a finite duration often called a transient change. In contrast to the traditional abrupt change detection, where the post-change period is assumed to be infinitely long, the detection of a suddenly arriving transient change should be done before it disappears. The detection of transient changes after their disappearance is considered as missed. Hence, the traditional quickest change detection criterion, minimizing the average detection delays provided a prescribed false alarm rate, is compromised. The proposed optimality criterion minimizes the worst case probability of missed detection provided that the worst case probability of false alarm during a certain period is upper bounded. A suboptimal CUSUM-type transient change detection algorithm, based on a subclass of truncated Sequential Probability Ratio Tests, is proposed. The optimization of the proposed algorithm in this subclass leads to a specially designed Finite Moving Average Test. The proposed method is analyzed theoretically and by simulation. A special attention is paid to the case of Gaussian observations with a dynamic profile.

14 citations


Journal ArticleDOI
TL;DR: This paper addresses the problem of distinguishing between two vector lines observed through noisy measurements and proposes a suboptimal test, called the epsilon most stringent test, which has a very simple form and its statistical properties are expressed in closed-form.
Abstract: This paper addresses the problem of distinguishing between two vector lines observed through noisy measurements. This is a hypothesis testing problem where the two hypotheses are composite since the signal amplitudes are deterministic and not known. An ideal criterion of optimality, namely the most stringent test, consists in minimizing the maximum shortcoming of the test subject to a constrained false alarm probability. The maximum shortcoming corresponds to the maximum gap between the power function of the test and the envelope power function which is defined as the supremum of the power over all tests satisfying the prescribed false alarm probability. The most stringent test is unfortunately intractable. Hence, a suboptimal test, called the epsilon most stringent test, is proposed. This test has a very simple form and its statistical properties are expressed in closed-form. It is numerically shown that the proposed test has a small loss of optimality and that it outperforms the generalized likelihood ratio test.

5 citations


Cites background or methods from "Constrained Epsilon-Minimax Test fo..."

  • ...The notations used throughout the paper come from [25], [26]....

    [...]

  • ...It can be concluded that (15) It must be noted that is a symmetric function, i.e., (16)...

    [...]


Journal ArticleDOI
TL;DR: The main contribution of the paper is the design of the Bayesian test with a quadratic loss function and its asymptotic study and the numerical experiments show that the proposed test outperforms theBayesian test associated to the 0—1 loss function when compared by using the quadratics loss function.
Abstract: The Bayesian test with 0—1 loss function is a standard solution to solve a multiple hypothesis testing problem in the Bayesian framework. For a large number of applications (like the intrusion detection, the anomaly detection,…) the alternative hypotheses have quite different importance and 0—1 loss function does not reflect the reality. The quadratic loss function can be more appropriate to distinguish the concurrent hypotheses. The main contribution of the paper is the design of the Bayesian test with a quadratic loss function and its asymptotic study. When the signal-to-noise ratio tends to infinity, it is theoretically established that the error probabilities of the proposed test coincide with the error probabilities of the standard one associated to the 0—1 loss function. In the non-asymptotic case, the numerical experiments show that the proposed test outperforms the Bayesian test associated to the 0—1 loss function when compared by using the quadratic loss function.

4 citations


Journal ArticleDOI
TL;DR: The optimization procedure for computing the discrete boxconstrained minimax classifier is presented, and a projected subgradient algorithm which computes the prior maximizing this concave multivariate piecewise affine function over a polyhedral domain is considered.
Abstract: In this paper, we present the optimization procedure for computing the discrete boxconstrained minimax classifier introduced in [1, 2]. Our approach processes discrete or beforehand discretized features. A box-constrained region defines some bounds for each class proportion independently. The box-constrained minimax classifier is obtained from the computation of the least favorable prior which maximizes the minimum empirical risk of error over the box-constrained region. After studying the discrete empirical Bayes risk over the probabilistic simplex, we consider a projected subgradient algorithm which computes the prior maximizing this concave multivariate piecewise affine function over a polyhedral domain. The convergence of our algorithm is established.

1 citations


Proceedings ArticleDOI
02 Dec 2017
TL;DR: A Bayesian test with a modified quadratic loss function is proposed to solve a multiple hypothesis testing (MHT) problem and the conditional asymptotic equivalence between these two tests is theoretically established.
Abstract: A Bayesian test has been previously proposed for a multiple hypothesis testing (MHT) problem with a quadratic loss function such that this problem can fit with some real applications where the concurrent hypotheses should be distinguished. However, this MHT problem as well as this quadratic loss function are insufficient for some other applications such as the simultaneous intrusion detection and localization in a wireless sensor network (WSN). This kind of applications could be considered as a MHT problem with null hypothesis. Therefore, a Bayesian test with a modified quadratic loss function is proposed to solve this MHT problem. The non-asymptotic bounds for analyzing the performance of the proposed test and the Bayesian test with the 0-1 loss function are obtained, from which the conditional asymptotic equivalence between these two tests is then theoretically established. The effectiveness of these bounds and the analysis on the conditional asymptotic equivalence are verified by the simulation results.

1 citations


Cites background from "Constrained Epsilon-Minimax Test fo..."

  • ...MOTIVATION AND CONTRIBUTION This paper deals with a multiple hypothesis testing (MHT) problem [1][2][3] in the Bayesian framework....

    [...]

  • ...Remark 1: The MHT problem with an unknown prior distribution has been tackled in a minimax framework [2][5][6]....

    [...]


References
More filters

Book
01 Jan 1983

25,004 citations


Book
01 Jan 1976
Abstract: (1977). Graph Theory with Applications. Journal of the Operational Research Society: Vol. 28, Volume 28, issue 1, pp. 237-238.

7,219 citations


Book
01 May 1997
TL;DR: Gaph Teory Fourth Edition is standard textbook of modern graph theory which covers the core material of the subject with concise yet reliably complete proofs, while offering glimpses of more advanced methods in each chapter by one or two deeper results.
Abstract: Gaph Teory Fourth Edition Th is standard textbook of modern graph theory, now in its fourth edition, combines the authority of a classic with the engaging freshness of style that is the hallmark of active mathematics. It covers the core material of the subject with concise yet reliably complete proofs, while offering glimpses of more advanced methods in each fi eld by one or two deeper results, again with proofs given in full detail.

6,244 citations


Book
01 Aug 1998
TL;DR: This self-contained and comprehensive book sets out the basic details of multiuser detection, starting with simple examples and progressing to state-of-the-art applications.
Abstract: From the Publisher: The development of multiuser detection techniques is one of the most important recent advances in communications technology. This self-contained and comprehensive book sets out the basic details of multiuser detection, starting with simple examples and progressing to state-of-the-art applications. The only prerequisites assumed are undergraduate-level probability, linear algebra, and digital communications. The book contains over 240 exercises and will be a suitable textbook for electrical engineering students. It will also be an ideal self-study guide for practicing engineers, as well as a valuable reference volume for researchers in communications, information theory, and signal processing.

5,047 citations


"Constrained Epsilon-Minimax Test fo..." refers background or methods in this paper

  • ...of user is the diagonal matrix of user amplitudes and is the vector whose th component is the antipodal symbol, or 1, transmitted by user [29], [30]....

    [...]

  • ...In order to make the simulation free of secondary details, let us consider the plainest, yet general enough, model of Direct-Sequence/Code-Division Multiple-Access (DS/CDMA) involving real signatures and Binary Phase Shift Keying (BPSK) data transmission [29], [30]....

    [...]

  • ...Proof: For a large value of , it is well-known that [29]...

    [...]


Journal ArticleDOI
Abstract: The paper reviews the state of the art of fault detection and isolation in automatic processes using analytical redundancy, and presents some new results. It outlines the principles and most important techniques of model-based residual generation using parameter identification and state estimation methods with emphasis upon the latest attempts to achieve robustness with respect to modelling errors. A solution to the fundamental problem of robust fault detection, providing the maximum achievable robustness by decoupling the effects of faults from each other and from the effects of modelling errors, is given. This approach not only completes the theory but is also of great importance for practical applications. For the case where the prerequisites for complete decoupling are not given, two approximate solutions—one in the time domain and one in the frequency domain—are presented, and the crossconnections to earlier approaches are evidenced. The resulting observer schemes for robust instrument fault detection, component fault detection, and actuator fault detection are briefly discussed. Finally, the basic scheme of fault diagnosis using a combination of analytical and knowledge-based redundancy is outlined.

3,215 citations


Frequently Asked Questions (1)
Q1. What are the contributions mentioned in the paper "Constrained epsilon-minimax test for simultaneous detection and classification" ?

In this paper, a constrained epsilon-minimax test is proposed to detect and classify non-orthogonal vectors in Gaussian noise, with a general covariance matrix, and in presence of linear interferences.