Pier Luigi Conti

Blind maximum likelihood estimation of traffic matrices under long-range dependent traffic

A new method, based on the maximum likelihood principle, through the numerical Expectation-Maximization algorithm, is proposed to estimate traffic matrices when traffic exhibits long-range dependence. The methods proposed so far in the literature do not account for long-range dependence. The method proposed in the present paper also provides an estimate of the Hurst parameter. Simulation results show that: (i) the estimate of the traffic matrix is more efficient than those obtained via existing techniques; (ii) the estimation error of the traffic matrix is lower for larger values of the true traffic intensity; (iii) the estimate of the Hurst parameter is slightly negatively biased.

Detection of anomalous bids in procurement auctions

Procurement auctions may be affected by abnormally low bids, whose acceptance may have negative consequences on the auctioneer. A method, based on the average submitted bid, is considered to detect such anomalous bids and aid the auctioneer in the possible rejection decision. Analytical expressions or simulation results are provided for the detection probability and for the false alarm probability. The performances heavily depend on the number of tenderers and on the dispersion of bid values. Both performance indices improve as the number of tenderers grows and generally degrade as the dispersion grows. The presence of multiple anomalous bids leads to a significant worsening of the performance, while courtesy bids raise both the false alarm probability and the detection probability. The use of the average-bid criterion, though officially endorsed in national legislations, is therefore recommended as a strongly precautionary criterion, i.e. when the need to avoid anomalous bids is considered much more relevant than the costs associated to deeper investigation of anomalous bids or to the erroneous rejection of regular bids.

Evaluation of matching noise for imputation techniques based on nonparametric local linear regression estimators

A new matching procedure based on imputing missing data by means of a local linear estimator of the underlying population regression function (that is assumed not necessarily linear) is introduced. Such a procedure is compared to other traditional approaches, more precisely hot deck methods as well as methods based on kNN estimators. The relationship between the variables of interest is assumed not necessarily linear. Performance is measured by the matching noise given by the discrepancy between the distribution generating genuine data and the distribution generating imputed values.

Uncertainty analysis for statistical matching of ordered categorical variables

The aim is to analyze the uncertainty in statistical matching for ordered categorical variables. Uncertainty in statistical matching consists in estimating a joint distribution by observing only samples from its marginals. Unless very restrictive conditions are met, observed data do not identify the joint distribution to be estimated, and this is the reason of uncertainty. The notion of uncertainty is first formally introduced, and a measure of uncertainty is then proposed. Moreover, the reduction of uncertainty in the statistical model due to the introduction of logical constraints is investigated and evaluated via simulation.

A rank-and-compare algorithm to detect abnormally low bids in procurement auctions

Detecting abnormally low bids in procurement auctions is a recognized problem, since their acceptance could result in the winner not being able to provide the service or work awarded by the auction, which is a significant risk for the auctioneer. A rank-and-compare algorithm is considered to detect such anomalous bids and help auctioneers in achieving an effective rejection decision. Analytical expressions and simulation results are provided for the detection probability, as well as for the false alarm probability. The suggested range of application of the detection algorithm leaves out the cases of many tenderers (more than 20) and quite dispersed bids (coefficient of variation larger than 0.15). An increase in the number of tenderers leads to contrasting effects, since both the false alarm probability and the detection probability are reduced. If the bids are spread over a large range, we have instead a double negative effect, with more false alarms and less detections. The presence of multiple anomalous bids worsens the performance of the algorithm as well. On the other hand, the method is quite robust to the presence of courtesy bids.

Asymptotic inference on a general measure of monotone dependence

In this paper a class of measures of monotone dependence (concordance/discordance) for arbitrary (not necessarily continuous) bivariate distributions is considered. It is shown that the corresponding sampling index of concordance/discordance (which is the most natural estimator of the population index) converges in law to a normal distribution. A Berry-Esseen bound for its rate of convergence is given. Finally, a consistent estimator of the asymptotic variance of the sampling concordance/ discordance index is proposed. This last result is essential for constructing confidence intervals and testing hypotheses on the population measure of monotone dependence.

Estimation of traffic matrices in the presence of long memory traffic

The estimation of traffic matrices in a communications network on the basis of a set of traffic measurements on the network links is a well-known problem, for which a number of solutions have been proposed when the traffic does not show dependence over time, as in the case of the Poisson process. However, extensive measurements campaigns conducted on IP networks have shown that the traffic exhibits long range dependence. Here a method is proposed for the estimation of traffic matrices in the case of long range dependence, and its theoretical properties are studied. Its merits are then evaluated via a simulation study. Finally, an application to real data is provided.

On a procedure to test whether the random variables of a sequence are independent and identically distributed, with applications to telephone and packet-switched networks

Abstract In this paper a new class of test-statistics is proposed. This class of statistical tests is particularly useful to test the null hypothesis that an arrival process is a Poisson process, and widely applicable to teletraffic problems, including telephone and packet-switched networks. After an introduction where a wide set of possible applications is examined (Section 1), in Sections 2 and 3 the class of test-statistics under examination is introduced, and its main theoretical properties are studied. In Section 4 a numerical evaluation (based on Monte Carlo simulation methods) of the power of the test is performed. Finally, in Section 5 an application to real data is considered.

How far from identifiability? A systematic overview of the statistical matching problem in a non parametric framework

ABSTRACT Statistical matching consists in estimating the joint characteristics of two variables observed in two distinct and independent sample surveys, respectively. In a parametric setup, ranges of estimates for non identifiable parameters are the only estimable items, unless restrictive assumptions on the probabilistic relationship between the non jointly observed variables are imposed. These ranges correspond to the uncertainty due to the absence of joint observations on the pair of variables of interest. The aim of this paper is to analyze the uncertainty in statistical matching in a non parametric setting. A measure of uncertainty is introduced, and its properties studied: this measure studies the “intrinsic” association between the pair of variables, which is constant and equal to 1/6 whatever the form of the marginal distribution functions of the two variables when knowledge on the pair of variables is the only one available in the two samples. This measure becomes useful in the context of the reduction of uncertainty due to further knowledge than data themselves, as in the case of structural zeros. In this case the proposed measure detects how the introduction of further knowledge shrinks the intrinsic uncertainty from 1/6 to smaller values, zero being the case of no uncertainty. Sampling properties of the uncertainty measure and of the bounds of the uncertainty intervals are also proved.

Nonparametric evaluation of matching noise

In this paper, the difference between the data generating process and the imputation procedures used in statistical matching is evaluated. The investigated imputation procedures are the distance hot deck and those referring to the kNN method, both with fixed and variable number of donors k. The matching noise is evaluated formally and investigated via a simulation.