This paper outlines my PhD project, which focuses on using artificial intelligence to classify and extract semantic information from breast radiology reports, specifically within the Algerian healthcare context. Currently, there are no systems in Algeria that automate the classification and processing of these reports. Our goal is to develop a system that not only addresses local needs but also compares favorably with systems from other countries and languages. The project explores tasks such as automatic parsing, classification, summarization, recommendation, and extraction, providing healthcare professionals with tools to manage a large number of reports more efficiently in a short time. Preliminary results are promising, and further work is focused on enhancing these results using large datasets and covering all NLP tasks that can assist in diagnosing breast cancer, with plans to expand to other medical documents and diseases.

Cancer pathology reports are integral to cancer diagnosis and treatment. The growing volume of these reports necessitates automated classification, prompting a shift towards rule-based Natural Language Processing (NLP) systems due to their capacity for more contextualized processing. This review examines these rule-based approaches, highlighting the key features extracted from the pathology reports, and evaluating the benefits they offer. Despite certain limitations, rule-based methods demonstrate considerable promise in the classification of cancer pathology reports. We conclude by identifying future research opportunities aimed at addressing these limitations to further enhance the benefits of the classification process.

The study of climate variation strongly depends on data availability and reliability which is collected from multiple sources in different types (multimodal). Up-today It was difficult to have a homogeneous climatic database for Algeria, since some climatic parameters (insolation, temperature, relative humidity and rainfall) are measured over several years and sometimes with data gaps without treatment or combination in many geographical regions. This work proposes a new multi-level data fusion approach to study the climate variation from multi-source and multimodal parameters on spatial and temporal vectors in Algeria. It facilitates the localization of geographical zones with high climate variation potential through fusion of classifier's decision results in high-level stage. A new Linked data model has been developed to manage links from multi-source as well as to build data fusion rules at the low-level stage.

Tremendous efforts have been put into the development of Feature Selection (FS) methods by the machine learning community. In this paper, we present basics surrounding this topic, providing its general process, evaluation procedure and metrics. In addition, we thoroughly discuss major application aspects. We also provide a comprehensive overview and comparison of some existing feature selection methods. We conclude this work by highlighting some critics, challenges, and future research directions of feature selection.

User's interests are important in query reformation, filtering or recommender systems. Besides, the user profile is always dynamic due to the instability of his interests and the increase among documents. Thus, its evolution is required however identifying the changes in user's interests over time may be challenging. Recent works focused on user interests evolution, takes into account the collection by matching the document and the query or by analyzing log file of user's interactions with the system (feedbacks). However, they did not consider the queries exclusively in the evolution process. This paper proposes a new user profile evolution approach in which user's queries and interests are used to update the user profile. To do so, we use genetic algorithm technique to extract relevant interests so that to detect new interests, improve the weights of the existing ones or delete the useless among them from the profile.

The exponential use of digital cameras has raised a new problem: how to store/retrieve images/albums in very large photo databases that correspond to special events. In this paper, we propose a new probabilistic graphical model (PGM) to recognize events in photo albums stored by users. The PGM combines high-level image features consisting of scenes and objects detected in images. To consider the discriminative power of features, our model integrates the object/scene relevance for more precise prediction of semantic events in photo albums. Experimental results carried out on the challenging PEC dataset with 807 photo albums are presented.

The Internet of Things (IoT) is a rapidly growing technology in recent years. It represents an extension of the Internet into the physical world embracing everyday objects. In consequence, users' privacy and security are becoming a great challenge. To cope with this issue, sophisticated approaches are needed to guarantee these services and hence ensure a large-scale adoption of IoT. In our work we present an approach to model and describe the architecture of internet of things, in which user privacy and security are ensured while taking into consideration the dynamic context in which evolve users, their experience and preferences with respect to security and privacy.

The International Charter1"Space and Major Disasters" is regularly activated during a catastrophic event and offers rescue teams comprehensive damage maps. Most of these maps are built by means of satellite image manual processing, which is often complex and demanding in terms of time and energy. Automatic processing supplies prompt treatment. Nevertheless, it usually presents a semantic gap handicap. The exploitation of ontologies to bridge the semantic gap has been widely recommended due to their quality of knowledge representation, expression, and discovery. In this work, we present an ontology-based semantic hierarchical classification method to undertake this problem. Ontology components are translated to image-based parameters and exploited to assist the classification process at two levels, and using 12 classes. The region of interest is selected from the first level and exhaustively analyzed and classified at the second level. The 2010 Haiti earthquake was selected as study area for this work. Experiments were performed using very high resolution multi-temporal QuickBird imagery and eCognition software.

Microblogging sites have become a staple in our modern world. They provide the users with the ability to keep in touch with their contacts, using up of 140 characters in the case of Twitter sites. Responding to this emerging trend, it becomes critically important to interactively view and analyze the massive amount of microblogging data from different perspectives and with multiple granularities. In the area of Business intelligence, On-line analytical processing (OLAP) is a powerful primitive for data analysis. However, OLAP tools face major challenges in manipulating unstructured text such as microblogging data.

Traditional data warehousing technologies and On-Line Analytical Processing (OLAP) are unable to analyze textual data. Moreover, as OLAP queries of a decision-maker are generally related to a context, contextual information must be taken into account during the exploitation of data warehouses. Thus, we propose a contextual text cube model denoted CXT-Cube which considers several contextual factors during the OLAP analysis in order to better consider the contextual information associated with textual data. CXT-Cube is characterized by several contextual dimensions, each one related to a contextual factor. In addition, we extend our aggregation OLAP operator for textual data ORank (OLAP-Rank) to consider all the contextual factors defined in our CXT-Cube model. To validate our model, we perform an experimental study and the preliminary results show the importance of our approach for integrating textual data into a data warehouse and improving the decision-making.