Real-time Air Pollution prediction model based on Spatiotemporal Big data

A new study introduces a Long Short-Term Memory (LSTM) model designed to predict NBA game outcomes using a comprehensive dataset spanning from the 2004-05 to 2024-25 seasons. This model utilizes an extensive sequence of 9,840 games to effectively capture evolving team dynamics and dependencies across seasons, addressing challenges faced by traditional prediction models.

A recent study has introduced a hybrid framework that integrates Long Short-Term Memory (LSTM) networks with the Merton-Lévy jump-diffusion model for financial forecasting. This approach utilizes the Grey Wolf Optimizer for hyperparameter tuning and evaluates its predictive performance against benchmark models using real-world datasets, including Brent oil prices and the STOXX 600 index.

A new method combining persistent homology analysis with discrete Fourier transform has been introduced for image compression, allowing for effective feature extraction from noisy datasets. This approach enables the differentiation of meaningful data while achieving compression levels comparable to JPEG across six metrics.

An innovative AI-powered Autonomous Underwater Vehicle (AUV) system has been developed to enhance sea exploration and scientific research, addressing challenges such as extreme conditions and limited visibility. The system utilizes advanced technologies including YOLOv12 Nano for real-time object detection and a Large Language Model (GPT-4o Mini) for generating structured reports on underwater findings.

A recent study highlights the issue of data leakage in Long Short-Term Memory (LSTM) networks used for time series forecasting, revealing that improper sequence construction before dataset partitioning can lead to misleading evaluation results. The research evaluates three validation techniques under both leaky and clean conditions, demonstrating how validation design can influence leakage sensitivity and performance metrics such as RMSE Gain.

A novel approach named PRIMRose has been introduced to predict energy values for individual residues in proteins with double InDel mutations, utilizing a Convolutional Neural Network architecture. This method focuses on the localized energetic impacts of mutations rather than global energy shifts, providing deeper insights into structural and functional disruptions caused by these mutations.