http://dspace.dtu.ac.in:8080/jspui/handle/repository/20350 Dissertation 2026-04-28T04:03:32Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22058 Title: A COMPARATIVE STUDY ON MACHINE LEARNING BASED STOCK PREDICTION BY INCORPORATING SENTIMENT ANALYSIS USING FINBERT Authors: SINGH, VIBHOR Abstract: Swift evolution of ML techniques, the financial forecasting scenario have been transformed tremendously. The domain of stock market prediction is one of the prominent applications of ML considering its inherent complexity and economic implication. Time series and statistical model was conventional cornerstone of market analysis, but they have limited potential for capturing intricate pattern. NLP sentiment analysis played a game changing role for more efficient stock prediction. NLP taps into huge pool of textual data of different sources, comprising of news and social media outlets. The collective mood and opinion of market participants can now be harness using power of NLP sentiment analysis which is to be fed to predictive model for better forecasting. Fusion of sentiment derived insights with ML algorithms presents a substantial leap which not only surges the predictive power of existing models but also provide nuanced understanding of the psychology of market movements driving factors. Consequently, financial industry witnessing a paradigm shift for the anticipation of stock prices fluctuations, with the support of AI driven sentiment analysis This paper presents a machine learning-based stock prediction model that integrates sentiment analysis using FinBERT, it is a specialized model for financial sentiment analysis that uses BERT. This study focuses on enhancing financial stock forecasting by adding investors sentiment data with conventional stock price data. This study takes into consideration traditional time series model like SARIMA for stock price prediction and three FinBERT infused ML models namely SVR, RFR, GBR. Eventually all predictive models are compared through regression evaluation metrics like MAE, MSE, RMSE, R2. 2025-05-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22046 Title: OPTIMISING ENERGY EFFICIENCY IN WIRELESS SENSOR NETWORKS: A STUDY OF CLUSTER-BASED HIERARCHICAL ARCHITECTURE Authors: SHARDA Abstract: Wireless sensor networks (WSNs) and IoT (Internet of Things) involve numerous small sensor nodes (SNs) running on batteries, expending energy in data routing towards the sink. An efficient routing scheme is crucial for WSN longevity. In WSN routing, minimising routing hops conflicts with inter-hop routing distance - fewer hops may increase distance and vice versa. To address this, we introduce a new Multi-Objective Biogeography Based Optimization algorithm (MOBBO-R) to optimize both routing objectives and reduce energy consumption in WSNs. MOBBO-R seeks a Pareto optimal solution, prolonging WSN lifetime by enhancing routing efficiency. Validated through MATLAB, MOBBO-R outperforms past routing algorithms like PSO routing and neighbourhood routing by approximately 7% and 22%, respectively. This algorithm holds promise for various IoT and IoE-based applications globally. 2024-06-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22045 Title: STUDY AND COMPARATIVE ANALYSIS OF DEEP LEARNING TECHNIQUES FOR EMOTION DETECTION IN TEXTUAL DATA Authors: BALIYAN, PURVAK Abstract: Detecting emotions from textual data is increasingly significant in various domains such as mental health support, customer experience enhancement, and social media analysis. As digital communication grows, understanding the emotional tone of written content has become essential for building responsive and human-centric applications. This study explores recent developments in deep learning techniques that aim to improve the accuracy and efficiency of emotion classification in text. In-depth analysis is conducted on a variety of deep learning architectures from basic ones like Bi-LSTM, Bi-GRU, ANN, CNN to complex transformer-based models like BERT, RoBERTa, and GPT. The models are evaluated on three widely used datasets—ISEAR, GoEmotion, and MELD— chosen for their diversity in the use of language as well as emotion categories. For the sake of comparison, the three datasets went through the same preprocessing including text cleaning, normalization, tokenization, and encoding. The aim of this research work is to analyze and compare the performance of the models under uniform training conditions and metric parameters like accuracy and F1-score. Experimentation outcomes show that transformer models provide enhanced performance by efficiently comprehending the contextually suitable sense of emotion in words. This comparative review not only unveils the potential of existing models but also reveals where more can be gained, the keys to even more sophisticated and sensitive emotion recognition systems. 2025-05-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22044 Title: ANALYSIS AND DEVELOPMENT OF TEXT-TO-SQL TRANSLATION SYSTEM USING LARGE LANGUAGE MODELS (LLMs) Authors: SHUKLA, PRADYUMN Abstract: The increasing reliance on data-driven decision-making has brought intuitive database access into limelight, particularly for inexperienced users. Text-to-SQL technologies bridge this shortcoming by converting natural language queries to SQL queries and thereby render database interaction more intuitive. Large Language Models have also influenced the Text2SQL system paradigm towards predicting correct and context-aware SQL. This survey maps the historical development of Text2SQL approaches from rule-based systems to LLM- based neural models. Extensive efforts have gone into using prompt engineering, schema alignment methods, and domain fine-tuning to ensure higher accuracy and generality. The models now exhibit significant progress in understanding complex queries as well as precise SQL code generation through emergent Large Language Model capabilities. The early systems had extremely strong template-based or rule-based mechanisms, whereas generation these days is extremely advanced neural systems brimming with domain knowledge and highly specialized embeddings. Although LLMs, particularly GPT and BERT, have really set the bar high for query interpretability and execution accuracy, there are significant challenges regarding meeting the needs of domain specificity, intricate queries, and scalability across heterogeneous schemas. It also pointed out how the RAG generation mechanism has been integrated and called for a paradigm shift towards adopting TAG for richer schema interaction. Future directions involve developing explainable models, fine-tuning multiturn conversational capabilities, and optimizing computational efficiency toward robustness and ease of use for Text2SQL systems. By addressing the gaps, the study lays the foundations for innovations in database querying, using LLMs to redefine accessibility and usability for a wide range of users. 2025-05-01T00:00:00Z