Papers by Bouchra Lamrini
Physics/Applied physics Detection of functional states by the 'LAMDA' classification technique: application to a coagulation process in drinking water treatment
The present Note proposes a learning classification methodology to identify functional states on ... more The present Note proposes a learning classification methodology to identify functional states on a coagulation process involved in drinking water treatment. In this work, we chose to carry out the supervised control of this process while using the LAMDA (Learning Algorithm for Multivariate Data Analysis) classification technique. The LAMDA classification technique proposes the interactive participation of the expert operator during the learning phase and in the optimisation of the classification. In this work, all information stemming from the environment process as well as expert knowledge has been aggregated and exploited. The application chosen for state identification is the Rocade drinking water treatment plant located at Marrakech, Morocco. To cite this
Le travail presente dans le cadre de cette etude vise a exploiter les modeles de deteriorations d... more Le travail presente dans le cadre de cette etude vise a exploiter les modeles de deteriorations des chaussees rigides integres au sein du logiciel HDM-4. Notre objectif est de caler experimentalement ces modeles en fonction de la deterioration de la chaussee sur la RP3606, region El Gara a la province de Settat realisee en 1990. La methode de calage des modeles de comportement adoptee au sein de la HDM-4 consiste dans un premier temps a comparer les defauts predits et observes par rapport aux valeurs observees pour la section temoin (section El Gara). Autrement, il s'agit d'etablir des facteurs d'ajustement notes K xxx qui permettent de minimiser l'erreur entre le modele cale et les valeurs des defauts observes sur cette section. Cette methode est basee sur un traitement des donnees en utilisant la regression lineaire simple ou la variable dependante est celle observee lors du releve visuel effectue par le CNER. La valeur independante est celle predite par le modele ...
Quand la Préservation passe par la Classification : Le Cas des Documents Sonores et Musicaux
ABSTRACT
Le document numérique est présent partout dans les activités de nos sociétés contemporaines où le... more Le document numérique est présent partout dans les activités de nos sociétés contemporaines où les échanges médiatisés s'accroissent. Les tâches de constitution et de classification des documents numériques deviennent complexes, notamment dans un flux numérique de gestion électronique de documents où les outils informatiques disponibles ne sont pas toujours adéquats. Nous proposons dans cet article, une approche d'analyse et de classification des patchs de la musique interactive. Soumis aux difficultés de la préservation, ces processus temps réel de traitement sonore sont considérés souvent comme des véritables documents numériques, ils sont à la fois supports de création et supports de constitution de connaissances dans la création artistique contemporaine. En s'appuyant sur l'analyse des expressions algébriques contenues dans des documents transcrits à partir des processus existants, conçus dans l'environnement Max/MSP, un lien systématique entre classification conceptuelle et classification par apprentissage automatique est donc établi afin d'établir les prémices d'une organologie des traitements musicaux et audio numériques.
Revue des sciences de l'eau, 2007
Résumé Le travail présenté propose une méthodologie de classification par apprentissage qui perme... more Résumé Le travail présenté propose une méthodologie de classification par apprentissage qui permet l’identification des états fonctionnels sur une unité de coagulation impliquée dans le traitement des eaux de surface. La supervision et le diagnostic de ce procédé ont été réalisés en utilisant la méthode de classification LAMDA (Learning Algorithm for Multivariate Data Analysis). Cette méthodologie d’apprentissage et d’expertise permet d’exploiter et d’agréger toutes les informations provenant du procédé et de son environnement ainsi que les connaissances de l’expert. L’étude montre qu’il est possible d’ajouter aux informations issues des capteurs classiques (température, matières en suspension, pH, conductivité, oxygène dissous), la valeur de la dose de coagulant calculée par un capteur logiciel développé dans une étude antérieure afin d’affiner le diagnostic. Le site d’application choisi pour l’identification des états fonctionnels est la station de production d’eau potable Rocade ...
A neural network system for modelling of coagulant dosage used in drinking water treatment
Adaptive and Natural Computing Algorithms, 2005
This paper presents the elaboration and validation of “soft sensor” using neural networks for on-... more This paper presents the elaboration and validation of “soft sensor” using neural networks for on-line estimation of the coagulation dose from raw water characteristics. The main parameters influencing the coagulant dosage are firstly determined via a PCA. A brief description of the methodology used for the synthesis of neural model is given and experimental results are included. The training of
Detection of functional states by the ‘LAMDA’ classification technique: application to a coagulation process in drinking water treatment
Comptes Rendus Physique, 2005
... There are two selectors that may bechanged by the user only when unsupervised learning has be... more ... There are two selectors that may bechanged by the user only when unsupervised learning has been chosen. These are the maximum desired variationpercentage and the maximum allowed iterations number. ... Lamrini et al. ...
In this paper∗, we investigate an unsupervised machine learning method based on one-class Support... more In this paper∗, we investigate an unsupervised machine learning method based on one-class Support Vector Machines for anomaly detection in network traffic. In the absence of any prior expert knowledge on anomalous data, we propose the use of a similarity measure for Multivariate Time Series to evaluate the output results and select the best model. A set of Key Performance Indicators, oriented for network and traffic monitoring, has been used to demonstrate the promising performance of the unsupervised learning approach.
Food Control, 2012
This paper presents a dynamic model of the kneading process based on artificial neural networks. ... more This paper presents a dynamic model of the kneading process based on artificial neural networks. This dynamic neuronal model allows predicting the bread dough temperature and the delivered power necessary to carry out mechanical work. This neuronal technique offers the advantage of very short computational times and the ability to describe nonlinear relationships, sometimes causal, explicit or implicit, between the input and output of a system. We used the recurrent neural networks to capture the dynamic of the process. The type and the number of inputs to the neural networks, as well as the nature of the learning set, the architecture and the parameter learning technique have been studied. The comparison of the results with experimental data shows the possibility to predict the temperature and the power delivered to the dough for various operating conditions.
The aim of this work is intended to research a direct approach providing from an experimental dat... more The aim of this work is intended to research a direct approach providing from an experimental database the optimal operating conditions of a process. To validate and elucidate, the methodology is implemented on colour evolution during beans pistachios roasting. The colour desired is defined by three attributes L-(light or dark), a(redness/green), and b-(blue/yellow). The fruits and oilseeds roasting is the most important stage in the quality development of the finished product, revealing the organoleptic characteristics of nuts and beans transformed. It is a complex technological process during which various reactions involved and thus modifying the colour, the flavour and the texture of the end product. This work presents the approach development based on artificial neural networks to model the impact of roasting process on colour pistachios. Computations of colour defining the colouring increment compared to the initial data are also proposed. The neuronal system developed present...
Desalination and Water Treatment, Sep 19, 2018
Drinking water is vital for everyday life. We are dependent on water for everything from cooking ... more Drinking water is vital for everyday life. We are dependent on water for everything from cooking to sanitation. Without water, it is estimated that the average, healthy human won't live more than 3-5 days. The water is therefore essential for the productivity of our community. The water treatment process (WTP) may vary slightly at different locations, depending on the technology of the plant and the water it needs to process, but the basic principles are largely the same. As the WTP is complex, traditional laboratory methods and mathematical models have limitations to optimize this type of operations. These pose challenges for water-sanitation services and research community. To overcome this matter, deep learning is used as an alternative to provide various solutions in WTP optimization. Compared to traditional machine learning methods and because of its practicability, deep learning has a strong learning ability to better use data sets for data mining and knowledge extraction. The aim of this survey is to review the existing advanced approaches of deep learning and their applications in WTP especially in coagulation control and monitoring. Besides, we also discuss the limitations and prospects of deep learning.
Data Mining - Methods, Applications and Systems [Working Title]
Among the learning algorithms, one of the most popular and easiest to understand is the decision ... more Among the learning algorithms, one of the most popular and easiest to understand is the decision tree induction. The popularity of this method is related to three nice characteristics: interpretability, efficiency, and flexibility. Decision tree can be used for both classification and regression kind of problem. Automatic learning of a decision tree is characterised by the fact that it uses logic and mathematics to generate rules instead of selecting them based on intuition and subjectivity. In this review, we present essential steps to understand the fundamental concepts and mathematics behind decision tree from training to building. We study criteria and pruning algorithms, which have been proposed to control complexity and optimize decision tree performance. A discussion around several works and tools will be exposed to analyze the techniques of variance reduction, which do not improve or change the representation bias of decision tree. We chose Pima Indians Diabetes dataset to cover essential questions to understand pruning process. The paper's original contribution is to provide an up-to-date overview that is fully focused on implemented algorithms to build and optimize decision trees. This contributes to evolve future developments of decision tree induction.
Uploads
Papers by Bouchra Lamrini