It was established in 1995 as a member of the group Seth Ved Parkash Mukand Lal Educational society. JMIT is affiliated with Kurukshetra University, in Kurukshetra. It is accredited by the All India Council for Technical Education, a division of the state Ministry of Human Resource Development. It was also accredited by NBA.
The frequency of the emergence of a cloudburst has increased significantly and the rate of clouds is measured up to 100 millimeters per hour which has made these phenomena the dominant and apex destructive hydrometeorological hazards in the hilly regions of India. In 2025 alone, the region in the upper northern mountainous zone, i.e., Uttarakhand, Himachal Pradesh, and Jammu & Kashmir, experienced around 765 incidents of extreme heavy rainfall, which is the highest that was recorded in the last five years. Uttarakhand recorded about 25 direct instances of cloud burst occurrences from July to August 2025, causing flash flooding and landslides that made cause discomfort to hydropower infrastructure/transportation networks. Between midmountain June and September 2025 Himachal Pradesh alone reported 266 deaths and economic losses to the tune of Rs 4080 crore of which at least 17 are proven to have died because of the cloudburst incidences. In order to overcome the imbalance, we hereby propose a hybrid CNN-BiLSTM-Attention architecture with Bayesian probabilistic output layer. The model makes use of multimodal inputs, based on INSAT-3D/3DR satellite imagery, Doppler radar mosaics and topographic covariance and ERA5 reanalysis data. Over time and with input BiLSTM component learning temporal dynamics of storm evolution and CNN, spatial convection structures are unraveled and attention mechanism highlights the high intensity cores of prevailing rainfall. Using Monte Carlo dropout ensembles improves the interpretability using SHAP and Grad-CAM analysis.
We present a carbon-aware fraud detection pipeline for historical fraud benchmarks using efficient tabular models (logistic regression, LightGBM, XGBoost) with SHAPguided feature selection and interpretability. Experiments on IEEE-CIS (2017–2018) and CreditCard (2013) datasets employ CodeCarbon emissions tracking with stratified cross-validation as the primary validation protocol. LightGBM achieves strong performance–emissions trade-offs: IEEE-CIS PR-AUC 0.740±0.002 (2 seeds), CreditCard PR-AUC 0.876 (single seed); emissions 8.6±0.6 mg CO₂e (IEEE) and 0.48 mg CO₂e (CreditCard). Runtime profiling reveals SHAP overhead dominates (96% of total emissions). We provide methodological frameworks for time-stratified validation, cost-sensitive optimization, SHAP-free deployment (projected 96% emissions reduction), and early stopping strategies. Comprehensive reproducibility artifacts, carbon-aware + SHAP-free + early stopping and production deployment extrapolations support sustainable fraud detection research.
Cotton is a vital agricultural crop that significantly contributes to the global textile industry; however, its productivity is severely affected by various leaf diseases caused by insects, bacteria, fungi, and viruses. Early and accurate detection of these diseases is essential to minimize crop losses and ensure sustainable farming practices. In this study, a hybrid framework based on Deep Learning is proposed for multi-level classification and detection of cotton leaf diseases. The framework integrates Convolutional Neural Networks (CNNs) with advanced techniques such as attention mechanisms and hybrid learning models to effectively extract and learn discriminative features from leaf images. The system performs hierarchical classification, including binary classification (healthy vs diseased), multiclass disease identification, and severity level estimation, along with localization of infected regions using object detection methods. The experimental results demonstrate the effectiveness of the proposed model, achieving an overall accuracy of 96.6%, precision of 95.7%, recall of 96.8%, and F1-score of 96.2% across multiple disease classes. The model shows strong performance in identifying diseases such as aphids, bacterial blight, and leaf curl virus under varying conditions. The integration of preprocessing, data augmentation, and hybrid modeling enhances the robustness and generalization capability of the system. The proposed approach provides a reliable and scalable solution for realtime cotton disease detection, supporting precision agriculture and enabling timely decision-making for improved crop health and yield.
Climate change threatens global food security through shifts in temperature and precipitation regimes, greater frequency of extreme weather events, and cascading effects on agricultural input markets and institutional capacity. Policymakers require nationally comparable diagnostic tools that are reproducible, transparent, and grounded in open data. This study presents an interpretable machine learning framework for country-level crop yield prediction and climate sensitivity screening, using publicly available FAOSTAT-derived panel data spanning 101 countries and 10 staple crop types over 1990–2013. A gradient-boosted decision tree model (XGBoost 2.1.4) is trained on observations from 1990 to 2008 and evaluated on a strictly held-out temporal window (2009–2013), using annual mean temperature, annual precipitation, total pesticide use (as a rough proxy for agricultural input management intensity), crop type and country identifiers, and temporally lagged yield values as predictive features. The optimized model yields high predictive accuracy on held-out data (R2 = 0.982; RMSE = 11,183 hg/ha; MAE = 4396 hg/ha). Ablation analysis reveals that model performance depends primarily on temporal yield persistence and crop identity, with performance declining to R2 = 0.940 when lagged features are omitted, while climate-only variables explain limited variation (R2 = 0.119). Notably, an ordinary least squares (OLS) baseline achieves comparable performance (R2 = 0.984), suggesting that the dominant predictive signals arise from a stable temporal–cross-sectional structure rather than nonlinear modeling flexibility. SHAP-based feature attribution identifies regime-dependent temperature effects, with larger (more negative) marginal contributions under high-temperature conditions. Stylized sensitivity perturbations (+1 °C, +2 °C, −20% pesticide inputs) indicate modest mean yield changes (−1.3% to −1.6%) but substantial cross-national heterogeneity. Systematic residual analysis identifies countries exhibiting consistent over- or under-prediction patterns, offering diagnostic signals for further institutional investigation. This framework is designed as a transparent, scalable screening tool for evidence-based prioritization rather than a validated causal or forecasting instrument. It complements localized agronomic expertise and supports SDG 2 (Zero Hunger) and SDG 13 (Climate Action).
This chapter presents a bibliometric analysis of 1,380 peer-reviewed articles on sustainable product promotion through green marketing strategies and e-commerce platforms published between 2010 and 2023. The study uses data from Web of Science, Scopus, and selected Google Scholar articles, analyzed through VOSviewer and the Bibliometrix R-package for co-citation analysis, keyword co-occurrence mapping, and intellectual structure visualization. Findings reveal exponential growth in publications after 2018 due to increasing environmental awareness, platform governance requirements, and sustainability concerns accelerated by the pandemic. Six major thematic clusters emerged: consumer behavioral antecedents, sustainability governance on digital platforms, transparency in the value chain, effectiveness of green advertising, regulatory compliance, and emerging technology applications. The study also identifies research gaps such as limited geographical diversity and overreliance on survey-based methodologies.