Journal of Big Data Technology and Business Analytics

AI-Based Chatbots: A Comprehensive Study of Architecture, Applications, Performance, Challenges, Ethics, and Future Perspectives

Suraj R. Nalawade, H. O. Tapase, Shreya Jadhav — Tue, 15 Sep 2026 00:00:00 +0000

Artificial intelligence (AI) chatbots powered by natural language processing (NLP) have transformed human-computer interaction across sectors such as e-commerce, healthcare, and customer service. This paper reviews the evolution of chatbot technology, with a particular focus on the components that constitute modern systems, including NLP engines, dialogue management systems, and backend integrations. It examines rule-based and AI-driven chatbot models, comparing their capabilities and limitations, and reports a performance evaluation covering intent-recognition accuracy, response time, task completion rate, and user satisfaction. The paper further addresses persistent technical challenges, including ambiguity in user queries, context retention across multi-turn conversations, and language variability, alongside ethical concerns such as data privacy and algorithmic bias. Emerging directions, including emotional intelligence, multimodal interaction, and real-time multilingual support, are also discussed. The findings indicate that while chatbots have achieved substantial gains in usability and efficiency, sustained progress depends on more robust context management and on responsible, transparent design practices.

Adaptive AI Models for Real-Time IoT Data Analytics: Frameworks, Concept Drift Mitigation, and Edge Implementation

Ashwini Kumbhar, Akshaya Uttekar, Dattatraya Kumbhar — Mon, 21 Sep 2026 00:00:00 +0000

The rapid proliferation of Internet of Things (IoT) devices has generated unprecedented volumes of high-velocity streaming data. Traditional, static machine learning (ML) models often experience significant performance degradation over time due to concept drift dynamic shifts in statistical data distributions caused by environmental changes, sensor degradation, or evolving operational conditions. This paper investigates adaptive artificial intelligence (AI) models tailored for real-time IoT analytics. By combining incremental online learning algorithms, lightweight drift-detection mechanisms, and hybrid edge-cloud architectures, the proposed framework maintains high predictive accuracy while meeting low-latency and resource constraints. Experimental evaluations across industrial and smart-environment benchmark datasets demonstrate that adaptive models achieve up to 96.4% classification accuracy under dynamic streaming conditions with a sub-50 ms detection latency. Furthermore, key security vulnerabilities and data management challenges inherent to decentralized stream processing are analyzed alongside differential privacy and federated mitigations.

Stock Market Prediction with Machine Learning using Fundamental Analysis: A Review

Singh RaviPratap Anand, Aakash Parmar — Sat, 03 Oct 2026 00:00:00 +0000

Stock market prediction has moved much beyond short-term technical trading signals toward deeper, value-oriented approaches grounded in fundamental analysis. This review synthesizes the findings of the previous article and studies that explore how machine learning models will utilize financial statements and accounting ratios and macroeconomic indicators to forecast stock returns, trends, and risk. Across the literature, several algorithms-including Random Forest, Support Vector Regression, Logistic Regression, Gradient Boosting techniques, and deep learning models such as LSTM and GRU-have been applied to fundamental data. The available proofs suggest that models built on fundamental indicators can perform better than those relying only on technical features, especially in long-term forecasting. Hybrid approaches that combine fundamental signals with technical signals or sentiment data further improve predictive robustness. Tree-based ensemble models are widely known for their interpretability and stability, while the deep learning architectures display strength in capturing temporal patterns in multi-year financial data, albeit with higher computational requirements and overfitting risks. Despite good outcomes, common challenges remain: limited data availability, quarterly reporting constraints, sector-specific variations, and insufficient integration of real-time macroeconomic factors. In general, this review of the literature highlights that integrating machine learning with structured fundamental analysis provides a reliable foundation for long-term stock prediction and informed portfolio decision-making, while highlighting the need for scalable, interpretable, and hybrid modeling frameworks for real-world financial applications.