Thanks to continuously evolving data management solutions, data-driven strategies are considered the main success factor in many domains. These strategies consider data as the backbone, allowing advanced data analytics. However, in the agricultural field, and especially in fish farming, data-driven strategies have yet to be widely adopted. This research paper aims to demystify the situation of the fish farming domain in general by shedding light on big data generated in fish farms. The purpose is to propose a dedicated data lake functional architecture and extend it to a technical architecture to initiate a fish farming data-driven strategy. The research opted for an exploratory study to explore the existing big data technologies and to propose an architecture applicable to the fish farming data-driven strategy. The paper provides a review of how big data technologies offer multiple advantages for decision making and enabling prediction use cases. It also highlights different big data technologies and their use. Finally, the paper presents the proposed architecture to initiate a data-driven strategy in the fish farming domain.

Deep learning (DL)-oriented document processing is widely used in different fields for extraction, recognition, and classification processes from raw corpus of data. The article examines the application of deep learning approaches, based on different neural network methods, including Gated Recurrent Unit (GRU), long short-term memory (LSTM), and convolutional neural networks (CNNs). The compared models were combined with two different word embedding techniques, namely: Bidirectional Encoder Representations from Transformers (BERT) and Gensim Word2Vec. The models are designed to evaluate the performance of architectures based on neural network techniques for the classification of CVs of Moroccan engineering students at ENSAK (National School of Applied Sciences of Kenitra, Ibn Tofail University). The used dataset included CVs collected from engineering students at ENSAK in 2023 for a project on the employability of Moroccan engineers in which new approaches were applied, especially machine learning, deep learning, and big data. Accordingly, 867 resumes were collected from five specialties of study (Electrical Engineering (ELE), Networks and Systems Telecommunications (NST), Computer Engineering (CE), Automotive Mechatronics Engineering (AutoMec), Industrial Engineering (Indus)). The results showed that the proposed models based on the BERT embedding approach had more accuracy compared to models based on the Gensim Word2Vec embedding approach. Accordingly, the CNN-GRU/BERT model achieved slightly better accuracy with 0.9351 compared to other hybrid models. On the other hand, single learning models also have good metrics, especially based on BERT embedding architectures, where CNN has the best accuracy with 0.9188.

The VUML (View-based UML) profile makes it possible to create unified, multiview models for complex systems. Nevertheless, the work already done focuses mostly on structural issues and does not fully cover behavioral concerns. With an emphasis on multi-view objects, this study adds capabilities to the VUML profile to represent system behavior. We suggest modifying UML state machines to express view object behaviors and how they combine to form global multi-view object behaviors. Our method allows for the independent construction of view behaviors, followed by a methodical composition process. The outcomes demonstrate how our approach improves VUML's capacity to represent intricate, multi-view system behaviors while preserving concern separation. The work being done here helps to close the gap that exists between structural and behavioral modeling in multi-view systems, which has implications for developing complex software systems using model-driven development.