Ousmane Sall

Followers

Following

Public Views

Steven Pinker

Harvard University

Steffen Boehm

University of Exeter

Michele Brun

University of Cagliari

Armando Marques-Guedes

UNL - New University of Lisbon

Kamel MEHDI

ESSTT

Fabio Cuzzolin

Oxford Brookes University

J.-C. Spender

Kozminski University

Isad Saric

University of Sarajevo

Viacheslav Kuleshov

Stockholm University

Umit Unver

Yalova University

Interests

Uploads

Papers by Ousmane Sall

COMFO : Corpus Multilingue pour la Fouille d’Opinions

L'utilisation d'algorithmes de Machine Learning (ML) en fouille d'opinions notamment ceux d'appre... more L'utilisation d'algorithmes de Machine Learning (ML) en fouille d'opinions notamment ceux d'apprentissage supervisé nécessite un corpus annoté pour entrainer le modèle de classification afin de prédire des résultats proches de la réalité. Malheureusement, il n'existe pas encore de ressources pour le traitement automatique de données textuelles exprimées dans le langage urbain sénégalais. L'objectif de cet article est de construire un corpus multilingue pour la fouille d'opinions (COMFO). Le processus de constitution du corpus COMFO est composé de trois étapes à savoir la présentation de la source de données, la collecte et préparation de données, et l'annotation par approche lexicale. La particularité de COMFO réside dans l'intégration des langues étrangères (française et anglaises) et celles locales notamment le wolof urbain afin de refléter l'opinion collective des lecteurs sénégalais.

Download

Study of Supervised Learning Algorithms for the Prediction of Road Accident Severity in Senegal

2022 7th International Conference on Computational Intelligence and Applications (ICCIA)

Predicting the Severity of the Road Accidents in Senegal: An Empirical Study

2022 the 5th International Conference on Data Storage and Data Engineering (DSDE)

Road accidents are a major problem, the consequences of which have negative impacts on the econom... more Road accidents are a major problem, the consequences of which have negative impacts on the economy, health and development of a country in general. The recrudescence of road accidents in Senegal during these last years has aroused a great debate in which certain responsibilities have been pointed out. A serious road traffic accident constitutes in most cases an emotional and economic disaster, hence the need to find sustainable solutions for the improvement of road safety. The objective of this work is to identify the major risk factors of road accidents in Senegal in order to set up a training corpus for the predictive study of the gravity of these accidents. According to our research on the bibliography, there is no study on the implementation of a corpus adapted to the data of accidents in Senegal. In this research, we used an exploratory study on the collected data to determine the factors most likely to cause accidents and supervised learning for predictive modeling of accident severity. The results showed that modelling with supervised learning can be used as an accident severity prediction system with promising accuracy.

Download

State of the art on mechanisms for personalization of information

Proceedings of the 7th International Conference on Software Engineering and New Technologies, 2018

With the exponential growth in the volume of data and the heterogeneity of information, users are... more With the exponential growth in the volume of data and the heterogeneity of information, users are facing increasing difficulties in identifying relevant information, irrelevant information, or even noise. To overcome these issues, mechanisms for customizing information have emerged. The personalization of information consists in adapting the information according to the preferences and/or access context of the user. Context can be seen as any information that describes an entity. The user's preferences are combined in a profile. The user profile can be seen as a setting in factor of the invariant part of the user's preferences. The mechanisms for personalizing information are mainly based on the user profile and the user's access context. The personalization of information cannot be seen as a domain in itself because it is used in various technological fields, namely: the field of information retrieval, the field of databases and the field of human-machine interfaces. The objective of this work is to provide an overview of the mechanisms of information personalization by presenting their foundations and showing you how information personalization is used in the various technological areas mentioned above.

FactExtract: Automatic Collection and Aggregation of Articles and Journalistic Factual Claims from Online Newspaper

2018 Fifth International Conference on Social Networks Analysis, Management and Security (SNAMS), 2018

Nowadays the web is very rich that's why it is the principal corpus of fact checkers. These j... more Nowadays the web is very rich that's why it is the principal corpus of fact checkers. These journalists in their quest for the truth daily scan manually a lot of sources in search of relevant informations. But in the data-journalism context where data are always increasing in real time, this manual exploration became impossible and more and more laborious. One of the actual used techniques for this problem is the automatic extraction of information on web pages more known as «web scraping ». The principal goal of web scraping is to retrieve on a web page. The main purpose of web scraping is to bring out a web page, specific and highly structured data with a reduced human effort. In this paper, we present an automatic extractor of articles and journalistic claims implemented on 15 Senegalese news websites.

Automation of Fact-Checking: State of the Art, Obstacles and Perspectives

2017 IEEE 15th Intl Conf on Dependable, Autonomic and Secure Computing, 15th Intl Conf on Pervasive Intelligence and Computing, 3rd Intl Conf on Big Data Intelligence and Computing and Cyber Science and Technology Congress(DASC/PiCom/DataCom/CyberSciTech), 2017

Today, the main problem in data journalism is quality. Indeed, even if the web is a broad data so... more Today, the main problem in data journalism is quality. Indeed, even if the web is a broad data source, it is essential to note that everything is not good to take and that it is always necessary to control good quality, relevance of data and reliability of sources. In this paper, we introduce a state of the art of the automation of the fact-checking as well as a diagnosis of obstacles for its automation in a journalistic context.

Part-Of-Speech Tagging in French: State-of-the-Art and Obstacles

2020 Seventh International Conference on Social Networks Analysis, Management and Security (SNAMS), 2020

POS Tagging (Part-Of-Speech Tagging) is considered as a basic step for automatic natural language... more POS Tagging (Part-Of-Speech Tagging) is considered as a basic step for automatic natural language processing. It consists in identifying the grammatical category of each word in a text. It is tricky for a machine to understand the grammatical form of words in some languages including French. These difficulties are related to the labeling of unknown words, the ambiguitsy of certain words, the understanding of many linguistic syntaxes etc. However, several approaches and tools have been proposed in the literature. In this article, we are exploring the Part-Of-Speech Tagging in French.

SenOpinion: A New Lexicon for Opinion Tagging in Senegalese News Comments

2020 15th Iberian Conference on Information Systems and Technologies (CISTI), 2020

The complexity of comments from the Senegalese on-line press is mainly due to ambiguity. This has... more The complexity of comments from the Senegalese on-line press is mainly due to ambiguity. This has led to customized abbreviations and the high presence of local languages. These factors make the current opinion mining tools ineffective. In view of this complexity, our objective is to suggest an opinion lexicon to process these types of data. This lexicon will take into account Wolof, French and urban language words and expressions. The significance of such a study is to provide our local languages with tools for automatic processing of natural languages.

News Comments Modeling for Opinion Mining: The Case of Senegalese Online Press

2020 International Conference on Advances in Computing and Communication Engineering (ICACCE), 2020

Opinion Mining is the processing textual contents from on-line discussions in order to highlight ... more Opinion Mining is the processing textual contents from on-line discussions in order to highlight the opinions of net surfers. In the process, the identification of the comments through a model and a formalism which encompass the methods and the data is critical step. Unfortunately, this step is generally neglected in most of the Sentiment analyses studies. The objective of this article is to address this challenging issue of modeling News comments. The relevance of such a study is to enable easy data collection and to avoid the algorithmic complexity of sentiment analysis methods.

Estimation and propagation of change effects in ontology based on graph properties

2017 IEEE International Conference on Agents (ICA), 2017

Ontology evolution refers to the process of modifying ontology in response to a change in its con... more Ontology evolution refers to the process of modifying ontology in response to a change in its conceptualization. This activity is joined to measuring inconsistencies and effects propagation on objects depending to ontology. In this paper, we propose a propagation estimation of ontological change effects in large ontology. The propagation approach uses the graph properties to estimate impact degree of a change in another entity of the ontology. The approach determines also the vulnerability rate of an ontological entity to a change in the community on ontology to which it belongs.

News Articles and Facts Modeling: Linked Data or Linked Fact

2020 International Conference on Advances in Computing and Communication Engineering (ICACCE), 2020

A press article is a collection of statements (opinions and/or allegations) that are a priori non... more A press article is a collection of statements (opinions and/or allegations) that are a priori non-factual (the facts), i.e. an opinion on a specific subject. This article seeks to address the challenging issue of modeling news articles and facts. The authors also discuss the possibilities of modeling networks of factual articles based on the RED Model and Linked Data.

Automatic Categorization of Press Articles through Learning: The Case of Senegalese Online Press

2019 IEEE International Conference on Big Data (Big Data), 2019

Nowadays, the study of online press has become an issue of phenomenal research. From articles col... more Nowadays, the study of online press has become an issue of phenomenal research. From articles collections and merging, opinion mining, artificial intelligence or automatic classification to fact-checking; researches are developing and opening new perspectives. The main objective is to pave the way for the journalistic consumption of the future. However, though the literature review mentions previous studies on online press articles classification, no work on automatic classification of press articles based on a theme has been carried out yet. In this article, we set out a supervised classifier of journalistic articles applied on Senegalese online press.

From Intelligent Crawling to Inclusive Fact-Checking: An End-to-End System

2021 Fifth International Conference On Intelligent Computing in Data Sciences (ICDS), 2021

In this article, we present an aggregation platform of journalistic contents based on an intellig... more In this article, we present an aggregation platform of journalistic contents based on an intelligent crawler of articles from Online Press, of an inclusive fact-checking approach and an Opinion Mining method. More than 300 000 press articles from more than 145 Senegalese online information websites have been collected, processed, analyzed, aggregated, stored, and classified by category and theme, thanks to Machine Learning. The primary objective in these researches is to provide Web surfers, mainly journalists and fact-checkers, with a unique platform aggregating all the data related to Senegalese current affairs with analysis options, display and fact-checking.

Comparative study of the mechanisms for personalization of information

Proceedings of the 9th International Conference on Information Systems and Technologies, 2019

The relevance of the information provided, its intelligibility and its adaptation to customer usa... more The relevance of the information provided, its intelligibility and its adaptation to customer usage and preferences are key factors in the rejection or success of an information system. Thus, in a context of information superabundance and heterogeneity, the implementation of mechanisms for personalizing information has become essential. Personalizing information means adapting it to In most information personalization mechanisms, there is agreement on the usefulness of having a user profile. The latter can be seen as a factoring of the invariant part of the user's preferences. There are two types of profiles: the statistical profile and the inferred profile. For the construction and updating of this profile, several approaches and techniques can be used. The personalization of information can appear in two modes: in query or recommendation, and the systems using it can be divided into three groups: recommendation systems, contextual access systems and custom meta-search engines....

An Architecture for Opinion Mining on Journalistic Comments: Case of the Senegalese Online Press

Comments from the Senegalese online press can create important opportunities for the socio-econom... more Comments from the Senegalese online press can create important opportunities for the socio-economic and political actors of our country. These are potentially promising data and useful sources of information. However, the complexity of these data sets no longer allows current methods of opinion mining to exploit this type of comments. This complexity is caused by ambiguous sentences, out-of-context comments and the use of terms borrowed from national languages. To avoid the risk of not reflecting the collective opinion of Senegalese readers, we are interested in proposing an architecture solely for the purpose of valorizing journalistic comments. The architecture will highlight a new solution to solving these types of problems.

Senegalese Road Crash Data Analysis Based On Clustering And Association Rule Mining

2021 4th International Conference on Data Storage and Data Engineering, 2021

This paper presents a study on road accident data in Senegal by identifying associations within a... more This paper presents a study on road accident data in Senegal by identifying associations within accident cases. We use unsupervised classification techniques such as clustering and association rule mining. Pre-processing of our data set revealed subgroup structures within the data. To reduce the association rule mining algorithm's search space, we used the k-modes clustering method as the main segmentation task on road accident data. Then, association rule mining helps identify the different circumstances associated with an accident in each group obtained by the k-modes algorithm. The results of the study show that to improve road safety, the authorities in charge of road transportation could orient their prevention policies towards three main aspects: vehicle fleet renewal, respect for the Highway Code, and prohibition of large vehicles circulation, such as trucks used for goods transportation at rush hours.

Web Scraping: State-of-the-Art and Areas of Application

2019 IEEE International Conference on Big Data (Big Data), 2019

Main objective of Web Scraping is to extract information from one or many websites and process it... more Main objective of Web Scraping is to extract information from one or many websites and process it into simple structures such as spreadsheets, database or CSV file. However, in addition to be a very complicated task, Web Scraping is resource and time consuming, mainly when it is carried out manually. Previous studies have developed several automated solutions. The purpose of this article is to revisit the different existing Web Scraping approaches, categories, and tools, but also its areas of application.

Crystal structures of two CuII compounds: catena-poly[[chloridocopper(II)]-μ-N-[ethoxy(pyridin-2-yl)methylidene]-N′-[oxido(pyridin-3-yl)methylidene]hydrazine-κ4 N,N′,O:N′′] and di-μ-chlorido-1:4κ2 Cl:Cl-2:3κ2 Cl:Cl-dichlorido-2κCl,4κCl-bis[μ3-ethoxy(pyridin-2-yl)methanolato-1:2:3κ3 O:N,O:O;1:3:4κ...

Towards an automation of the fact-checking in the journalistic web context

International Journal of Big Data Intelligence, 2019

Is Fact checking automatisable? Apparently, yes, since numerous moved forward noted in the search... more Is Fact checking automatisable? Apparently, yes, since numerous moved forward noted in the search and the analysis of digital data. However, this task which in priori seemed to be simple, turns out rather binding. Indeed, automate the check of facts combine and requires at the same time very advanced knowledge in analysis of data, in search of web data, the web technologies, in image processing and sometimes in automatic natural language processing (NLP). Nevertheless, the latter years, numerous researches are led to deepen such an analysis. In this article, having revisited the state of the art concerning the question, we identify and diagnose in detail the obstacles before concluding with an explanation of the methods.

A search engine dedicated to the environment around the Congo basin area: Model and Architecture

ICST Transactions on Ambient Systems, 2018

Considering the environmental field around the Congo Basin area (in Central Africa), many initiat... more Considering the environmental field around the Congo Basin area (in Central Africa), many initiatives implement and communicate on their activities via the internet, yet the information they provide remains very little exploited, even inaccessible. The data is scattered on the internet and it is quite difficult to find information in a fairly precise way. Despite the existence of Internet research services (search engines) developed to facilitate the search for information in the vast data network that is the Internet, there are still concerns about quality, and the relevance of information provided in as research results. In this context, we present in this paper a construction approach and the architecture of a search engine dedicated to the environment around the Congo Basin area. This paper develops a theoretical approach that uses scientific analysis and empirical approaches to conceptualize the optimization of the relevance of the results of a thematic search engine by aggregating tools. This approach is a compilation of ideas, methods and tools that, put together, will improve the relevance of the results of a thematic research.

Download

COMFO : Corpus Multilingue pour la Fouille d’Opinions

Download

Study of Supervised Learning Algorithms for the Prediction of Road Accident Severity in Senegal

2022 7th International Conference on Computational Intelligence and Applications (ICCIA)

Predicting the Severity of the Road Accidents in Senegal: An Empirical Study

2022 the 5th International Conference on Data Storage and Data Engineering (DSDE)

Download

State of the art on mechanisms for personalization of information

Proceedings of the 7th International Conference on Software Engineering and New Technologies, 2018

FactExtract: Automatic Collection and Aggregation of Articles and Journalistic Factual Claims from Online Newspaper

2018 Fifth International Conference on Social Networks Analysis, Management and Security (SNAMS), 2018

Automation of Fact-Checking: State of the Art, Obstacles and Perspectives

Part-Of-Speech Tagging in French: State-of-the-Art and Obstacles

2020 Seventh International Conference on Social Networks Analysis, Management and Security (SNAMS), 2020

SenOpinion: A New Lexicon for Opinion Tagging in Senegalese News Comments

2020 15th Iberian Conference on Information Systems and Technologies (CISTI), 2020

News Comments Modeling for Opinion Mining: The Case of Senegalese Online Press

2020 International Conference on Advances in Computing and Communication Engineering (ICACCE), 2020

Estimation and propagation of change effects in ontology based on graph properties

2017 IEEE International Conference on Agents (ICA), 2017

News Articles and Facts Modeling: Linked Data or Linked Fact

2020 International Conference on Advances in Computing and Communication Engineering (ICACCE), 2020

Automatic Categorization of Press Articles through Learning: The Case of Senegalese Online Press

2019 IEEE International Conference on Big Data (Big Data), 2019

From Intelligent Crawling to Inclusive Fact-Checking: An End-to-End System

2021 Fifth International Conference On Intelligent Computing in Data Sciences (ICDS), 2021

Comparative study of the mechanisms for personalization of information

Proceedings of the 9th International Conference on Information Systems and Technologies, 2019

An Architecture for Opinion Mining on Journalistic Comments: Case of the Senegalese Online Press

Senegalese Road Crash Data Analysis Based On Clustering And Association Rule Mining

2021 4th International Conference on Data Storage and Data Engineering, 2021

Web Scraping: State-of-the-Art and Areas of Application

2019 IEEE International Conference on Big Data (Big Data), 2019

Towards an automation of the fact-checking in the journalistic web context

International Journal of Big Data Intelligence, 2019

A search engine dedicated to the environment around the Congo basin area: Model and Architecture

ICST Transactions on Ambient Systems, 2018

Download

Ousmane Sall

Uploads

Papers by Ousmane Sall

Log In