Afif Masmoudi - Academia.edu (original) (raw)

Papers by Afif Masmoudi

Review of Pacific Basin Financial Markets and Policies, Nov 17, 2021

Duration and convexity are important measures in fixed-income portfolio management. In this paper... more Duration and convexity are important measures in fixed-income portfolio management. In this paper, we analyze this measure of the bonds by applying the beta model. The general usefulness of the beta probability distribution enhances its applicability in a wide range of reliability analyses, especially in the theory and practice of reliability management. We estimate the beta density function of the duration/convexity. This estimate is based on two important and simple models of short rates, namely, Vasicek and CIR (Cox, Ingersoll, and Ross CIR). The models are described and then their sensitivity of the models with respect to changes in the parameters is studied. We generate the stochastic interest rate on the duration and convexity model. The main results show that the beta probability distribution can be applied to model each phase of the risk function. This distribution approved its effectiveness, simplicity and flexibility. In this paper, we are interested in providing a decision-making tool for the manager in order to minimize the portfolio risk. It is helpful to have a model that is reasonably simple and suitable to different maturity of bonds. Also, it is widely used by investors for choosing bond portfolio immunization through the investment strategy. The finding also shows that the probability of risk measured by the reliability function is to highlight the relationship between duration/convexity and different risk levels. With these new results, this paper offers several implications for investors and risk management purposes.

Springer eBooks, Aug 9, 2014

Review of Pacific Basin Financial Markets and Policies, 2021

Applied Mathematics, 2014

Empirical study on the factors that induce jumps in interest rates in the euro area is still miss... more Empirical study on the factors that induce jumps in interest rates in the euro area is still missing. In this paper, maximum likelihood estimates of I-distribution parameters are extracted using as a first step, an original linear model. According to the contribution of ([1] [2]) in the case of developing a class of Poisson-Gaussian model, we try to enhance the predictive power of this model by distinguishing between a pure Gaussian and Poisson-Gaussian distributions. Such an empirical tool permits to optimizing results through a comparative analysis dealing with the fluctuation of the Euro-interbank offered rate and its statistical descriptive behaviour. The analytical and empirical methods try to evaluate the behavioural success of the ECB intervention in setting interest rates for different maturities. Jumps in euribor interest rate can mainly be linked to surprise decisions of the European Central Bank, and the too frequent meetings of the ECB before November 2001. Despite this special event that leads to a certain lack of predictability, other few day-of-week effects are modelled to prove eventual evidence of bond market overreaction. Empirical results prove that Mondays and Wednesdays are the preponderant days. Regarding monetary policy, negative surprises induce larger jumps than positive ones.

Arabian Journal for Science and Engineering, 2020

2019-nCoV is a virulent virus belonging to the coronavirus family that caused the new pneumonia (... more 2019-nCoV is a virulent virus belonging to the coronavirus family that caused the new pneumonia (COVID-19) which has spread internationally very rapidly and has become pandemic. In this research paper, we set forward a statistical model called SIR-Poisson that predicts the evolution and the global spread of infectious diseases. The proposed SIR-Poisson model is able to predict the range of the infected cases in a future period. More precisely, it is used to infer the transmission of the COVID-19 in the three Maghreb Central countries (Tunisia, Algeria, and Morocco). Using the SIR-Poisson model and based on daily reported disease data, since its emergence until end April 2020, we attempted to predict the future disease period over 60 days. The estimated average number of contacts by an infected individual with others was around 2 for Tunisia and 3 for Algeria and Morocco. Relying on inferred scenarios, although the pandemic situation would tend to decline, it has not ended. From this perspective, the risk of COVID-19 spreading still exists after the deconfinement act. It is necessary, therefore, to carry on the containment until the estimated infected number achieves 0.

Filomat, 2019

In this paper, we introduce finite mixture models with singular multivariate normal components. T... more In this paper, we introduce finite mixture models with singular multivariate normal components. These models are useful when the observed data involves collinearities, that is when the covariance matrices are singular. They are also useful when the covariance matrices are ill-conditioned. In the latter case, the classical approaches may lead to numerical instabilities and give inaccurate estimations. Hence, an extension of the Expectation Maximization algorithm, with complete proof, is proposed to derive the maximum likelihood estimators and cluster the data instances for mixtures of singular multivariate normal distributions. The accuracy of the proposed algorithm is then demonstrated on the grounds of several numerical experiments. Finally, we discuss the application of the proposed distribution to financial asset returns modeling and portfolio selection.

Communications in Statistics - Theory and Methods, 2014

ABSTRACT

Among existing Bayesian network (BN) parametrizations, conditional Gaussian are able to deal with... more Among existing Bayesian network (BN) parametrizations, conditional Gaussian are able to deal with discrete and continuous variables. Bayesian estimation of conditional Gaussian parameter needs to define several a priori parameters which are not easily understandable or interpretable for users. The approach we propose here is free from this priors definition. We use the Implicit estimation method which offers a substantial computational advantage for learning from observations without prior knowledge and thus provides a good alternative to Bayesian estimation when priors are missing. We illustrate the interest of such estimation method by first giving the Bayesian Expectation A Posteriori estimator (EAP) for conditional Gaussian parameters. We then describe the Implicit estimator for the same parameters. One experimental study is proposed in order to compare both approaches. * aidajarraya@yahoo.fr

Our work aims at developing or expliciting bridges between Bayesian Networks and Natural Exponent... more Our work aims at developing or expliciting bridges between Bayesian Networks and Natural Exponential Families, by proposing discrete exponential Bayesian networks as a generalization of usual discrete ones. In this paper, we illustrate the use of discrete exponential Bayesian networks for Bayesian structure learning and density estimation. Our goal is to empirically determine in which contexts these models can be a good alternative to usual Bayesian networks for density estimation.

International Conference on Tools with Artificial Intelligence, 2011

In this paper, we develop the notion of discrete exponential Bayesian network, parametrization of... more In this paper, we develop the notion of discrete exponential Bayesian network, parametrization of Bayesian networks using more general discrete quadratic exponential families instead of usual multinomial ones. We then introduce a family of prior distributions which generalizes the Dirichlet prior classically used with discrete Bayesian network. We develop the posterior distribution for our discrete exponential BN leading to bayesian estimations of the parameters of our models and one new scoring function extending the Bayesian Dirichlet score used for structure learning. These theoretical results are finally illustrated for Poisson and Negative Binomial BNs.

Procedia Computer Science, 2015

The use of Bayesian Networks (BNs) as classifiers in different fields of application has recently... more The use of Bayesian Networks (BNs) as classifiers in different fields of application has recently witnessed a noticeable growth. Yet, the Naïve Bayes application, and even the augmented Naïve Bayes, to classifier-structure learning, has been vulnerable to certain limits, which explains the practitioners resort to other more sophisticated types of algorithms. Consequently, the use of such algorithms has paved the way for raising the problem of super-exponential increase in computational complexity of the Bayesian classifier learning structure, with the increasing number of descriptive variables. In this context, the present work's major objective lies in setting up a further solution whereby a remedy can be conceived for the intricate algorithmic complexity imposed during the learning of Bayesian classifiers structure with the use of sophisticated algorithms.

The present paper presents a theoretical extension of our earlier work entitled"A comparativ... more The present paper presents a theoretical extension of our earlier work entitled"A comparative study of two models SV with MCMC algorithm" cited, Rev Quant Finan Acc (2012) 38:479-493 DOI 10.1007/s11156-011-0236-1 where we propose initially a mixture stochastic volatility model providing a tractable method for capturing certain market characteristics. To estimate the parameter of a mixture stochastic volatility model, we first

Lecture Notes in Computer Science, 2012

Nowadays, Bayesian Networks (BNs) have constituted one of the most complete, self-sustained and c... more Nowadays, Bayesian Networks (BNs) have constituted one of the most complete, self-sustained and coherent formalisms useful for knowledge acquisition, representation and application through computer systems. Yet, the learning of these BNs structures from data represents a problem classified at an NP-hard range of difficulty. As such, it has turned out to be the most exciting challenge in the learning machine area. In this context, the present work's major objective lies in setting up a further solution conceived to be a remedy for the intricate algorithmic complexity problems imposed during the learning of BNstructure through a massively-huge data backlog. Our present work has been constructed according to the following framework; on a first place, we are going to proceed by defining BNs and their related problems of structurelearning from data. We, then, go on to propose a novel heuristic designed to reduce the algorithmic complexity without engendering any loss of information. Ultimately, our conceived approach will be tested on a car diagnosis as well as on a Lymphography diagnosis data-bases, while our achieved results would be discussed, along with an exposition of our conducted work's interests as a closing step to this work.

Neurocomputing, 2014

Our work 1 aims at developing or expliciting bridges between Bayesian networks (BNs) and Natural ... more Our work 1 aims at developing or expliciting bridges between Bayesian networks (BNs) and Natural Exponential Families, by proposing discrete exponential Bayesian networks as a generalization of usual discrete ones. We introduce a family of prior distributions which generalizes the Dirichlet prior applied on discrete Bayesian networks, and then we determine the overall posterior distribution. Subsequently, we develop the Bayesian estimators of the parameters, and a new score function that extends the Bayesian Dirichlet score for BN structure learning. Our goal is to determine empirically in which contexts some of our discrete exponential BNs (Poisson deBNs) can be an effective alternative to usual BNs for density estimation.

Statistics & Probability Letters, 2010

In this paper, we show that the multinomial exponential families in a d-dimensional linear space ... more In this paper, we show that the multinomial exponential families in a d-dimensional linear space are characterized by the determinant of their covariance matrix, named generalized variance.

Review of Quantitative Finance and Accounting, 2012

This paper examines two asymmetric stochastic volatility models used to describe the volatility d... more This paper examines two asymmetric stochastic volatility models used to describe the volatility dependencies found in most financial returns. The first is the autoregressive stochastic volatility model with Student's t-distribution (ARSV-t), and the second is the basic SVOL of Jacquier et al. (J Bus Econ Stat 14:429-434, 1994). In order to estimate these models, our analysis is based on the Markov Chain Monte-Carlo (MCMC) method. Therefore, the technique used is a Metropolishastings (Hastings in Biometrika 57: [97][98][99][100][101][102][103][104][105][106][107][108][109] 1970), and the Gibbs sampler (Casella and George in The Am Stat 46: 167-174, 1992; Gelfand and smith in J Am Stat Assoc 85:398-409, 1990; Gilks and Wild in 41:337-348, 1992). The empirical results concerned on the Standard and Poor's 500 composite Index (S&P), CAC40, Nasdaq, Nikkei and DowJones stock price indexes reveal that the ARSV-t model provides a better performance than the SVOL model on the MSE and the maximum Likelihood function.

Journal of Theoretical Biology, 2008

We introduce here the concept of Implicit networks which provide, like Bayesian networks, a graph... more We introduce here the concept of Implicit networks which provide, like Bayesian networks, a graphical modelling framework that encodes the joint probability distribution for a set of random variables within a directed acyclic graph. We show that Implicit networks, when used in conjunction with appropriate statistical techniques, are very attractive for their ability to understand and analyze biological data. Particularly, we consider here the use of Implicit networks for causal inference in biomolecular pathways. In such pathways, an Implicit network encodes dependencies among variables (proteins, genes), can be trained to learn causal relationships (regulation, interaction) between them and then used to predict the biological response given the status of some key proteins or genes in the network. We show that Implicit networks offer efficient methodologies for learning from observations without prior knowledge and thus provide a good alternative to classical inference in Bayesian networks when priors are missing. We illustrate our approach by an application to simulated data for a simplified signal transduction pathway of the epidermal growth factor receptor (EGFR) protein.

Learning Bayesian Network structure from database is an NP-hard problem and still one of the most... more Learning Bayesian Network structure from database is an NP-hard problem and still one of the most exciting challenges in machine learning. Most of the widely used heuristics search for the (locally) optimal graphs by defining a score metric and employs a search strategy to identify the network structure having the maximum score. In this work, we propose a new score

Journal of Computational Biology, 2009

We summarize here the Implicit statistical inference approach as an alternative to Bayesian netwo... more We summarize here the Implicit statistical inference approach as an alternative to Bayesian networks and we give an effective iterative algorithm analogous to the Expectation Maximization algorithm to infer signal transduction network when the set of data is incomplete. We proved the convergence of our algorithm that we called Implicit algorithm and we apply it to simulated data for a simplified signal transduction pathway of the EGFR protein.