BSBrahmaleen Kaur Sidhu

Brahmaleen Kaur Sidhu

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Brahmaleen Kaur Sidhu has a total of 12 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A A Generative AI-Based Intelligent Advisory Framework for Continuous Soil Health Monitoring and Adaptive Nutrient Management

    by Jaspreet Kaur, Rakesh Singh, Brahmaleen Kaur Sidhu

    Soil health is a critical determinant of agricultural productivity and sustainability; however, many farming regions continue to face degradation due to imbalanced fertilizer use and limited access to timely diagnostics. Existing soil assessment approaches provide baseline insights but remain largely static and inadequate for continuous, field-specific decision-making. This paper presents SoilCare Agent, a generative AI-driven advisory framework for continuous soil health assessment and adaptive nutrient management. The system leverages structured soil data and dynamic field inputs to assess soil conditions and generate actionable insights. At its core, a Large Language Model (LLM) translates analyzed soil attributes and contextual factors into personalized, human-readable recommendations. An interactive query interface further enables farmers to submit natural language questions related to soil health and crop management, receiving context-aware responses in real-time and thereby enhancing accessibility and practical usability. The novelty of this work lies in integrating LLM-driven interactive advisory generation with dynamic, context-aware, and user-friendly decision support beyond conventional static recommendation systems. Experimental evaluation demonstrates the framework’s potential to improve fertilizer decision-making and support sustainable, data-driven precision agriculture. The proposed framework offers a scalable and user-centric solution for real-world soil health management.

  • Open Access

    Explainable Machine Learning for ICU Mortality Prediction: Comparing SHAP, LIME and Anchors on MIMIC-IV

    by Brahmaleen Kaur Sidhu

    Mortality prediction in the intensive care unit represents a high-stakes clinical task in which machine learning models have demonstrated substantial predictive advantages over conventional severity scoring systems. However, the opacity of high-performing gradient boosting models impedes clinical adoption, undermines regulatory compliance, and limits the actionability of risk predictions. This paper presents a rigorous technical investigation of three post-hoc explanation methods—TreeSHAP, ensemble LIME, and Anchors—implemented and systematically compared on XGBoost and LightGBM mortality prediction models trained on the MIMIC-IV critical care database, encompassing 48,372 ICU admissions and 64 clinical features. A multi-criteria evaluation framework assessing faithfulness, stability, and comprehensibility is developed and applied, with quantitative metrics including the Area Over the Perturbation Curve, Lipschitz continuity estimation, and explanation entropy. Experimental results demonstrate that TreeSHAP achieves superior faithfulness (AOPC: 0.276) and stability (Lipschitz estimate: 0.312) whilst generating explanations in 4 milliseconds per instance, supporting real-time clinical deployment. Anchors produce the most comprehensible explanations through compact if-then rules but exhibit lower faithfulness and limited patient coverage. Ensemble LIME occupies an intermediate position across all dimensions. These findings provide evidence-based guidance for explanation method selection in clinical AI deployment contexts subject to the transparency obligations of the EU Artificial Intelligence Act and GDPR Article 22.

  • Open Access

    A Hybrid Deep Reinforcement Learning Approach for Algorithmic Trading in Commodity Futures Markets

    by Baljinder Kaur, Brahmaleen K. Sidhu, Gurjit Singh Bhathal

    The financial markets have been affected extensively in recent years, due to algorithmic trading, predominantly concerning commodity futures. This research provides a novel hybrid approach combining classical machine learning algorithms (random forest) with modern deep reinforcement learning algorithms (deep Q‐network, proximal policy optimization, and soft actor–critic) for algorithmic trading in the commodity futures market. In contrast to prior investigations which mostly focus on equities, we applied this hybrid model to precious metals (gold) futures using a novel dataset of gold prices augmented with geopolitical risk indices, resulting in outstanding performance on financial risk metrics. The evaluation done suggests that the soft actor–critic agent modeled outshines conventional approaches (buy‐and‐hold and sell‐and‐hold) along with deep Q‐network and proximal policy optimization, in relation to cumulative profit, Sharpe ratio, maximum drawdown, and annualized volatility. Our results enhance the algorithmic trading methodology growth by supporting problem reduction commonly associated with classical trading techniques or approaches. This study provides significant value for traders and financial institutions intending to implement artificial intelligence in their cognitive trading strategies.


  • Design of Real Time Intelligent System for Women Safety

    by Priyanka Kohli, Kawaljeet Singh, Brahmaleen Kaur Sidhu

    Background: Even the technology has advanced a lot in these modern times, women's safety is still a massive issue. Women are not safe anywhere. Their safety is an essential and decisive issue in today’s world. They face extreme troubles when they are alone, travelling on lonely roads and at workplace. Education is essential for everyone and especially for women. It is a powerful weapon to end poverty. Government and NGOs have made many rules and regulations but the crime against women has not ended. There are many instances of harassment, abuse, eve-teasing, kidnapping, domestic violence and rape. Women are working equally as men but still their safety is a big challenge. With the boom in the IT and corporate sector, many women are now working and even during nightshifts. Objectives: A Real Time Intelligent System for Women Safety has been proposed to overcome these limitations. Methods: In this paper, many women's safety systems that were proposed or developed have been discussed. Major drawbacks in the current system, devices or applications have been highlighted. Latest technologies like machine learning, artificial intelligence or data science have been ignored. Results: The proposed system will be highly sophisticated with the latest techniques and women will feel safe in all situations. It will predict the safe route while moving alone. Conclusion: Moreover, it will record the evidence and will also suggest the rights of women.


  • Open Access

    Predictive Analytics Models for Commodity Market: A Literature Review

    by Baljinder Kaur, Brahmaleen Sidhu, Gurjit Singh Bhathal

    This research presents a literature review on the application of Artificial Intelligence (AI) and Machine Learning (ML) techniques in predictive analytics, with a main emphasis on the commodity market. Predictive analytics influences statistical models and Machine Learning approaches to analyze periodic data and predict future outcomes, providing substantial value across various domains such as finance, healthcare, retail, marketing, and commodities. The motive of this review is two-fold: (1) to survey AI-ML based approaches, techniques, and tools employed in predictive analytics, and (2) to inspect predictive analytics models precisely designed for the commodity market. The review integrates current methodologies, strategies, algorithms, and tools, while also surveying the types of commodities analyzed and their interdependencies. The major influence of this research lies in recognizing major gaps in existing studies, predominantly in model generalization, integration of heterogeneous data sources, and evaluation of real-world applicability. These gaps emphasize possibilities for future research and provide direction for developing more robust, accurate, and scalable predictive analytics frameworks. DOI: https://doi.org/10.17762/ijisae.v12i3.7958

  • A machine learning approach to software model refactoring

    by Brahmaleen Kaur Sidhu, Kawaljeet Singh, Neeraj Sharma

    Good software quality is a consequence of good design. Model refactoring counteracts erosion of the software design at an early stage in the software development project complying with the model-driven engineering paradigm. Traditional model refactoring approaches work at the surface level by using threshold values of model metrics as indicators of suboptimal design and carry out localized corrections. Through this paper, it is proposed that identifying design flaws at a higher level of granularity will save from the vicious cycle of small refactoring operations and their cascaded side-effects. The notion of functional decomposition, as an anomalous design tendency and a dominant cause of design, smells in object-oriented software, is introduced. It is suggested that refactoring operations targeted at signs of functional decomposition instead of atomic smells achieve substantial improvement in design within a concise quality assurance procedure. The idea is realized using a deep neural network that learns to recognize the presence of functional decomposition in UML models of object-oriented software. The presented approach uses data science methods to gain insight into multidimensional software design features and uses the experience gained to generalize subtle relationships among architectural components.