Models & Optimisation and Mathematical Analysis Journal


Description

This journal concerns both the national and international scientific community and will be primarily focusing on Models and Optimisation of Systems. Systems will be utilized in different applications for example, Web technologies, Information Systems, Decision Systems, Embedded Systems, Control-command Systems and Real-time Systems. Space of journal is also dedicated to mathematical analysis like functional spaces, polynomial computing etc. MOMAjournal reserves a space for research work in the field of the history of mathematics and the didactics of mathematics.


8

Volumes

8

Numéros

67

Articles


Automatic Control of Multiple Pseudo- Relevance Feedback Pass

Mokhtari Chakir,  Debakla Mohammed,  Meftah Boudjelal, 

Résumé: Automatic query expansion (AQE) based on Pseudo Relevance Feedback (PRF) is a useful technique for enhancing the effectiveness of information retrieval systems. In this article, we propose the study of the behavior of a set of features based on the query and the set of feedback documents in the aim to control automatically the convergence of the query expansion process to the pertinent set of documents. It is a new look for which is commonly known in IR community as prediction of query performance or query ambiguity. The results obtained show that is very interesting to use these features as good predictors for any PRF methods. Index Terms— Information retrieval, Automatic query expansion, pseudo-relevance feedback, query performance.

Mots clés: Information retrieval ; Automatic query expansion ; pseudo-relevance feedback ; query performance


Measurement and Social Network Analysis With Parallel Frequent Pattern Mining

محمد الأمين يرمس,  محمد رباح, 

Résumé: Social network analysis (‘SNA’) measures are a vital tool for understanding the behavior of networks and graphs. However, the huge amount of data generated by networks require new techniques to optimize the calculation. Many efficient algorithms based on graph calculation have been used in the last decades; most of them do not scale the large amount of data. In this paper, we propose a new approach based on parallel frequent pattern mining as an essential data-mining task, with a goal of calculating degree centralities. To take advantage of the computing power provided by HPCs, the use of a hybrid distribution of data (horizontal and vertical) seems necessary to reduce the computing time. The implementation show how is benefic to combine parallel and distributed programming techniques such as MPI with data mining tools.

Mots clés: Parallel Frequent Pattern Mining ; Social Network Analysis ; Degree Centralities ; HPC


An Empirical Study on the effect of weighting schemes and Machine Learning algorithms on the Arabic text Classification

Bennabi Rim Sakina,  Elberrichi Zakaria, 

Résumé: Nowadays, many applications use text classification to categorize different documents with predefined labels. However, most of the work focuses on the English language and a small number of studies focus on the Arabic language. The latter is widely used on the Internet and has a complex morphology that needs to be studied in different manners. In this context, we propose in this paper an empirical study on the effect of the use of different learning algorithms (SVM, Naive Bayes and KNN) and different weighting methods (TC, TF and TF.IDF) on Arabic textual classification . The goal of our work is to find the best combination that enhance the performance. The results show that SVM and TF.IDF combination offers the best accuracy and F-Measure (94%).

Mots clés: weighting schemes ; NB ; SVM ; KNN ; Arabic text classification


Use of differents Classifiers for Recognition of Fear Emotions in speech

هركوس هواري, 

Résumé: This work consists on the automatic recognition of the emotions in the speech, because it plays a very significant role in the communication. The automatic recognition of the emotions potentially had a broad application in the Human Machine Interaction. In this work emotional speech corpus in Algerian Dialect was created for parameters extraction. The selected parameters in our study are the prosodic (pitch, intensity and duration), the unvoiced frames, jitter, shimmer and cepstral parameters MFCCs (Mel-Frequency Cepstral Coefficients) to analyze the emotions of fear and neutral. These parameters will be used in the automatic recognition of the emotions. The system of recognition is based on the methods of classification KNN (K-Nearest Neighbor), SVM (Support Vector Machine) and ANN (Artificial Neural Network). The obtained results lead us to observe that the use of MFCCs parameters gives a very acceptable rate of emotion recognition.

Mots clés: Speech emotion ; Algerian Dialect ; prosodic ; MFCC ; KNN ; SVM ; ANN