scientific research has consistently sought to improve human life quality, with a particular emphasis on advancing healthcare in hospitals and clinics. This study focuses on the development of intelligent systems to improve healthcare by assisting with disease diagnosis and treatment recommendations. Given the precision necessary in the medical area, we look at advanced technologies, specifically deep learning (DL) algorithms that use neural networks (NN) to simulate complex decision-making processes. We present a detailed review of the most recent breakthroughs in Automatic Radiology Report Generation (ARRG) systems, covering the many frameworks and approaches utilized in their implementation, as well as evaluating the most often used datasets and metrics. Additionally, we investigate the relevance of large language models (LLMs) in improving the development and interpretation of radiological reports. A proposed framework, detailing the best practices, future research directions, and implementation strategies, is also presented. This study aims to offer a thorough understanding of ARRG systems, benefiting researchers and practitioners in the field.

This research paper addresses the challenge of detecting communities in sparse social graphs and presents a novel approach that leverages node importance and label propagation. The proposed method consists of three phases: initialization, label assignment, and filtering. In the initialization phase, we carefully identify and designate key nodes using their local information and associate them with different labels. Subsequently, in the label assignment phase, the assigned labels are propagated to neighboring nodes, which are organized in a multilevel manner, taking into account their relevance and significance. Through the filtering phase, we effectively eliminate irrelevant labels, enhancing the accuracy of community assignments and resulting in an optimized community structure. To assess the effectiveness of our approach, we conducted experiments on both real-world networks and synthetic networks. A comparative analysis was performed against several established community detection techniques from existing literature. The results clearly demonstrate that our proposed algorithm surpasses existing methods in terms of accuracy and efficiency.

Stock markets trading has risen as a critical challenge for artificial intelligence research. Such environments require artificial agents to coordinate and transfer their best experience to other agents. However, the strongest agents have been trained using expert capabilities or employing hand-crafted experts features. Notwithstanding these improvements, no previous single system has come near to mastering the trading environment.

Big Data presents promising technological and economical opportunities. In fact, it has become the raw material of production for many organizations. Data is available in large quantities, and it continues generating abundantly. However, not all the data will have valuable knowledge. Unreliable sources provide misleading and biased information, and even reliable sources could suffer from low data quality.

Community detection in social networks is an area, which has witnessed many studies in recent years, and therefore several algorithms have been proposed. The majority of these methods are based on the relationships among users (structural information) to identify communities in social networks. However, these methods take into account only the strength of connections among users, but they ignore the content information such as the topics shared by users. In this paper, we propose a method of community detection in social networks that combines the content and the structural information. To meet this end, first, we propose a new approach to detect the topics involved in social networks by exploit the statistical and semantic measures. Second, we divide users into different groups according to their topics of interest, and then we perform a static community detection algorithm to detect communities in each group of users. The experimental results on real life datasets have shown that our method finding extracts more meaningful communities in social networks, and improves the quality of communities from the perspective of topics and links.

In many companies data is used as a source for creating knowledge in their sphere of business. Therefore this paper presents a knowledge capitalization framework in big data context. Based on the technology derived from distributed systems, our research concerns the design and development of a knowledge engineering framework in big data context. It can be integrated in any knowledge management system. The proposed framework is based on four layers: we start by extracting hidden topics, using the LDA approach in batch processing to handle the complexity of multi knowledge domains and to keep the semantic relations between knowledge entities. Then we use clustering mechanisms to pick the best combination between topics from different sources. As a result, we get, in every distributed site, the related topics (knowledge), in order to facilitate the research and access, to get the useful knowledge in real time processing.

Ontology matching is one of the essential methodologies to overcome heterogeneity issues. Multiple knowledge-based and information systems perform ontology matching strategies to find correspondences between several ontologies for the purpose of discovering valuable information across various domains. The design and implementation of matching systems raises several challenges, especially, the matching accuracy and the performance issues. Accordingly, adapting the system to the requirements of Big Data era brings additional perspectives and challenges. Furthermore, to provide on-the-fly matching and in-time processing, the system must handle matching accuracy, runtime complexity and performance issues as an entire matching strategy. To this end, this paper presents a new hybrid ontology matching approach that benefit on one hand from the opportunities offered by parallel platforms, and on the other hand from ontology matching techniques, while applying a resource-based decomposition to improve the performance of the system.

Systematic review is the scientific process that provides reliable answers to a particular research question. There is a significant shift from using manual human approach to decision support tools that provides a semi-automated screening phase by reducing the required time and effort. Text classification is useful in determining the statistical significance level of association rules to reduce workload in the systematic review. Several approaches to generate a Rule set for rule based classifiers were proposed in the literature. In this paper, we show that statistic as well as semantic measures of a rule can be combined and effectively computed as a hybrid feature selection rule measure (HFSRM). Moreover, we propose a new algorithm called Rules7-hybrid feature selection (Rules7-HFSRM) by combining the classical algorithm Rules7 and the HFSRM and then used it on the systematic review problem. Our results show that our algorithm significantly outperforms the state-of-the-art benchmark algorithms in the systematic review context.

Finance ontology is, in most cases, manually addressed. This results in a tedious development process and error prone that delay their applicability. This is why there is a need of domain ontology learning methods that built the ontology automatically and without human intervention. However, in this learning process, the discovery of non-taxonomic relationships has been recognized as one of the most difficult problems. In this paper, we propose a new methodology for learning non-taxonomic relationships and building financial ontology from scratch. Our new technique is based on using and adapting Open Information Extraction algorithms to extract and label domain relations between concepts. To evaluate our new method effectiveness, we compare the extracted non-taxonomic relations of our algorithm with related works in the same finance corpus. The results showed that our system is more accurate and more effective.

Systematic review is the scientific process that provides reliable answers to a particular research question by interpreting the current pertinent literature. There is a significant shift from using manual human approach to decision support tools that provides a semi-automated screening phase by reducing the required time and effort to the group of experts. Most of proposed works apply supervised Machine Learning (ML) algorithms to infer exclusion and inclusion rules by observing a human screener. Unless, these techniques holds very little promise in study identification phase, because the rate of excluding citations erroneously still unreasonable. In this paper, we contribute to this line of works by proposing an alternative approach, not yet tested in this domain based on semantic rule-based classifiers. This approach involved applying a novel Hybrid Feature Selection Method (HFSM) within a Class Association Rules (CARs) algorithm. Experiments are conducted on a corpus resulting from an actual systematic review. The obtained results show that our algorithm outperforms the existing algorithms in the literature.

Actually, massive amount of data is created every day due to the proliferation of free tools such as blogs, social media networks, large-scale e-commerce, websites, etc. to create and to share information. Big Data management and integration has become the core of modern science. It brings benefits for scientific disciplines, business areas by creating a radical shift in the way of managing data. However, existing works lack the flexibility to deal with diverse and changing data sources and the scalability to cope with large streaming data. They just focus on one part of the whole process of big data management (e.g. Storage, Integration, Processing or Knowledge Extraction). For the first time in the literature, this paper proposes a complete novel multi-layer prototype for Big Data management in order to process all the issues of Big Data management from data level to the extracted knowledge exploitation, by introducing semantic technologies to bridge the gap of Big Data Integration by adding flexibility, scalability and richness.

Renewable Energy (RE) field provides significant new challenges for research in Intelligent Decision Support Systems (IDSS) since decision-making process requires more intelligent algorithms and mechanisms that can solve complex problems involving a large number of stakeholders in an uncertainty, dynamic, and distributed environment. In this paper, we present an intelligent RE-DSS framework to facilitate the decision-making process for the planning and designing intelligent Renewable Energy Management Systems. The framework offers a widespread adoption of more intelligent components in classical RE-DSS that will eventually lead to more efficient decision-making in all levels of RE projects management.

Finding the appropriate information and understanding to human research is a delicate task when dealing with an outstanding number of unstructured texts created daily. Hence the objective of clustering algorithms which are part of the powerful text mining tools. In this paper, we propose a novel text document clustering based on a new hybrid feature selection method that we call HFSM. This technique extracts statistical and semantic relevant terms to pilot the clustering mechanism. The experiments conducted on Reuters corpus demonstrate the practical aspects of our algorithm and show that it generates more accurate clustering than the one obtained by other existing algorithms.

The Multi-Criteria Recommender systems continue to be interesting and challenging problem. In this paper we will propose an approach for selection of relevant items in a RS based on multi-criteria ratings and a method of computing weights of criteria taken from Multi-criteria Decision Making (MCDM). This method proposes a correlation coefficient and standard deviation integrated approach for determining weight of criteria in multi-criteria recommender systems. We evaluated the proposed method on an example of movies recommendation. Our approach was compared to some other metrics used in Information Theoretic approach to illustrate its potential applications.

The PageRank algorithm, proposed by [Page et al., 1998] is used in the Google search engine to improve the results of requests by taking into account the link structure of the Web. PageRank give the same weight to all pages that is the surfer model is proposed using a uniform distribution. Richardson and Domingoshave proposed a more interesting and intelligent surfer model combining the link and content information in PageRank. Given a multiple term query, the surfer selects a term according to some probability distribution and uses that term to guide its behavior. We propose to improve this algorithm by using a simultaneous multiple-terms query model. Firstly we propose a measure of relevance of a page to a simultaneous multiple terms query. Then we develop our performed intelligent surfer model. To evaluate the performance, we have tested our algorithm on the "Moroccan ministry tourism's Web and show that the performance is superior to that obtained by the existing algorithms.