The phenomenon of big data has become a defining characteristic of the technological age. Consequently, new challenges arise, including the types of data and the scale of the information available; traditional relational databases frequently prove inadequate in addressing these demands, resulting in the proliferation of NoSQL databases. This research presents a comprehensive review of NoSQL-based systems and their methodologies for managing extensive data. Furthermore, this study examines the primary categories of NoSQL databases, their optimal attributes, and their strategies for addressing the contemporary "5Vs" of Big Data: volume, velocity, variety, value, and veracity. Additionally, the research highlights the opportunities presented by NoSQL technologies in emerging fields, such as the Internet of Things (IoT), artificial intelligence, and machine learning, alongside the associated risks, including trade-offs in consistency, security challenges, and the lack of standardization. By integrating classification, comparison, and analysis, this review offers a unified perspective that can aid researchers and practitioners in the effective selection and implementation of NoSQL solutions within Big Data contexts.

Book recommendation systems (BRSs) play a vital role in digital libraries, online bookstores, and e-learning platforms by assisting users in discovering relevant content from vast collections. Traditional methods, such as collaborative filtering (CF), content-based filtering (CBF), and hybrid techniques, have historically formed the foundation of BRSs; however, they suffer from limitations including the cold-start problem, data sparsity, and overspecialization. In recent years, deep learning–based approaches have emerged as powerful alternatives, leveraging architectures such as CNNs, RNNs, BERT, and Neural Collaborative Filtering (NCF) to capture complex user–item interactions and support multimodal integration. This survey is the first to systematically review book recommendation systems published between 2020 and February 2025, filling a critical gap left by earlier studies that did not comprehensively examine this recent period of accelerated research. The paper introduces a novel taxonomy of BRSs that classifies systems according to methodological foundations, approaches, datasets, and evaluation metrics, while also identifying recurring challenges and emerging trends. The findings reveal a clear methodological transition from similarity-driven approaches to neural representation learning, reflecting the increasing demand for intelligent, scalable, and adaptive solutions. Traditional methods, however, remain essential as baseline models for benchmarking and comparative evaluation.

The rapid growth of digital content has intensified the problem of information overload, making it challenging for users to access relevant resources. Recommender systems (RSs) address this issue by filtering data and providing suggestions, thereby improving decision-making and user satisfaction. This paper presents a comprehensive review of recommender systems (RSs), with particular emphasis on their methods, techniques, benefits, history, and applications. It examines traditional approaches, including collaborative filtering, content-based filtering, and hybrid strategies, before providing a classification of deep learning models in recommender systems and analyzing their impact on enhancing RS capabilities. In addition, the paper discusses evaluation methods used to assess recommendation performance and highlights their roles in measuring system effectiveness. Finally, it synthesizes the key challenges confronting recommender systems, including data sparsity, scalability, and cold-start issues.