This study introduces a sophisticated dual-denoising framework for high-frequency momentum strategies in China's ChiNext market, integrating wavelet-based temporal filtering with isolation forest cross-sectional anomaly detection. Utilizing a comprehensive dataset of over 2 million daily observations from January 2016 to November 2025, the results demonstrate that wavelet denoising achieves exceptional efficacy for turnover series with a mean Signal-to-Noise Ratio improvement of 10.7 dB, while isolation forest robustly identifies anomalous stocks characterized by excessive trading activity and distorted risk-return profiles. Empirical results reveal that linear models with wavelet denoising... More >

In industrial settings, unplanned machine downtime is a serious risk to profitability, operational effectiveness, and production. In order to predict machine breakdowns before they occur, this research offers a machine learning-based predictive maintenance framework that enables early prediction of machine downtime. The research is carried out using recorded data sets of industrial machines that operate according to various factors or reasons for downtime. Based on these values, prediction of downtime is possible. To guarantee data quality and consistency, several preprocessing techniques, such as imputation and normalization, were used on a dataset of 2,500 records and 16 features, ranging... More >

Traffic management is a crucial issue due to the massive increase in vehicles on the road. In order to improve traffic flow, reduce congestion, and ensure commuter safety at junctions, traffic light regulation is essential. This research aims to minimize delay and accident risk while maximizing vehicle throughput through intersections. In this paper, we introduced a new hybrid approach by combining various algorithms, which include a genetic algorithm (GA), Simulated Annealing (SA), and Particle Swarm Optimization (PSO) to reduce traffic delay at traffic lights. The Webster delay function has been used as a fitness function to evaluate the algorithmic performance. Our hybrid algorithm effici... More >

Maternal mortality in Bangladesh remains a critical public health challenge, with recent evidence indicating stagnation in mortality reduction despite expanded facility-based delivery and skilled birth attendance. Accurate identification of high-risk cases is essential to enable targeted intervention and resource allocation. This study develops an interpretable machine learning framework for maternal mortality prediction using the nationally representative Bangladesh Maternal Mortality Survey 2016 (BMMS-2016). A comprehensive data integration and feature engineering pipeline was implemented across demographic, socioeconomic, and maternal healthcare domains. Given the severe class imbalance i... More >

Rice plays a crucial role globally, as it is widely consumed across nations. Therefore, studying rice prices is vital, since fluctuations in the price can affect both its consumption and availability. This study analyzes time-series models using an international dataset. After preprocessing, the dataset comprises 71,856 samples and eight input features from six countries. The original dataset contained 300,816 rows and 23 columns. This study aims to predict rice inflation rates using time series models such as ARIMA, LSTM, and BiLSTM. The ARIMA model achieved the best combination of values (4,1,4)(0,0,0). Various statistical techniques that calculate inflation rates require expert knowledge... More >

Missing value imputation in multivariate time series is a critical challenge in the field of data mining. Although Transformer-based methods excel in modeling long-range dependencies, their inherent point-wise attention mechanisms often lack explicit modeling of local inductive biases in time series, making it difficult to effectively capture local smoothness and evolutionary trends. Furthermore, existing feature embedding strategies struggle to fully decouple the internal temporal evolution of variables from complex cross-variable dependencies. To address these limitations, this paper proposes a novel dual-stage imputation framework named M-SAITS. This framework innovatively introduces a de... More >

Deepfake media is growing rapidly and causing significant harm. Bad actors now use AI to create fake videos that appear increasingly realistic. Traditional detection tools often fail because they analyze audio or visual signals in isolation. This paper introduces an intelligent Deepfake Detection system that addresses this limitation through a novel Multi-Modal Dispersion Framework. The system identifies subtle inconsistencies by tracking how lip movements align with speech patterns. By projecting these features into a shared latent space, the model quantifies the semantic divergence between modalities. A transformer module then captures cross-modal context to detect fine-grained manipulatio... More >

This study investigates convolutional neural network (CNN) architectures for predicting steering angles in self-driving vehicles navigating unstructured roads, using road-facing image data. Two complementary experiments are conducted. First, the impact of three activation functions—Exponential Linear Unit (ELU), Rectified Linear Unit (ReLU), and Leaky ReLU—is evaluated on a baseline CNN model. Trained on 14,754 images and validated on 3,585 images, the model with ELU activation achieves the lowest validation mean squared error (MSE) compared to ReLU and Leaky ReLU, demonstrating superior convergence and generalization. Second, the effect of model complexity is examined using ELU activati... More >