Islamic Azad University, Lahijan Branch

ABSTRACT Accurately predicting stock prices remains a major challenge in financial analytics due to the complexity and noise inherent in market data. Feature selection plays a critical role in improving both computational efficiency and predictive performance. In this study, we introduce a novel hybrid framework that integrates metaheuristic feature‐selection algorithms with an enhanced transformer‐based prediction model fine‐tuned using temporal embedding and adaptive attention pruning. We evaluate and compare the effectiveness of three nature‐inspired metaheuristic algorithms: bat algorithm (BAT), gray wolf optimization (GWO), and beluga whale optimization (BWO) for selecting the most informative features from a time‐series stock dataset. After feature selection, the optimal subsets are fed into our modified transformer equipped with temporal embeddings and adaptive attention pruning. Extensive experiments conducted on the Bharat Heavy Electricals Limited (BHEL) dataset show that the proposed hybrid framework outperforms traditional methods in terms of predictive accuracy. Among the evaluated approaches, the combination of BWO and the fine‐tuned transformer achieves the best performance, yielding a Test RMSE of 0.0030 and a Test MAPE of 0.0108, demonstrating the superiority of BWO in identifying relevant features. This work provides a comprehensive comparative analysis of hybrid metaheuristic–deep learning models for stock price prediction and offers a foundation for integrating more explainable and scalable AI techniques into financial forecasting.

Multidrug-resistant uropathogenic Escherichia coli (UPEC) are a major cause of urinary tract infections, largely due to efflux pump overexpression, porin alterations, and increasing biofilm formation. To address these challenges, we developed bisabololoxide A-loaded PEG400-oleate polymeric nanoparticles (BAPNs) and evaluated their potential to enhance the efficacy of ciprofloxacin against ciprofloxacin-resistant isolates. BAPNs were synthesized and characterized by FTIR, TEM, FE-SEM, EDX, TGA, and zeta potential analysis. The synergistic effects of BAPNs and ciprofloxacin were evaluated in drug-resistant isolates. The inhibition of Biofilm formationwas quantified by crystal violet staining and gene expression changes in some genes related to efflux pumps, porins and biofilm formation were analyzed in two representative isolates by qPCR. BAPNs exhibited a spherical morphology with sizes ranging from 17 to 28 nm (TEM) and 30–43 nm (FE-SEM). Checkerboard analysis showed that the MIC of ciprofloxacin decreased by 2- to 32-fold in ciprofloxacin-resistant isolates and the ATCC 8739 strain when combined with BAPNs. Biofilm inhibition was significantly enhanced with the combined therapy compared to ciprofloxacin alone. qPCR analysis further revealed significant downregulation of the AcrAB-TolC efflux pump genes and the virulence genes fimH and hlyA, along with upregulation of the porin genes (ompC and ompF). Furthermore, molecular docking analysis confirmed the strong binding affinities of bisabololoxide A with virulence-associated proteins of E. coli. Our findings suggested that BAPNs may significantly potentiate ciprofloxacin activity by inhibiting efflux pump expression and biofilm formation, and increase of cell membrane permeability to drugs (with enhancing porins), thereby restoring susceptibility in resistant UPEC isolates. These results highlight BAPNs as a promising adjunctive nanocarrier strategy to overcome ciprofloxacin resistance in uropathogenic E. coli. Further validation in larger isolate panels and in vivo studies is warranted.

Gastric Cancer remains a significant global health challenge due to its high mortality rate and nonspecific early symptoms, which often lead to late diagnosis and poor prognosis. This study proposes an advanced machine learning approach for the early detection of gastric cancer using metabolomic data extracted from blood plasma. The method integrates Categorical Boosting with the Water Optimization Algorithm and Heptagonal Reinforcement Learning to enhance classification accuracy. The dataset, comprising metabolomic, clinical, and follow-up data from 702 plasma samples, was preprocessed and balanced using the Statistical Deviation Mode Balancer. Feature selection was performed using Adaptive Lasso to identify key biomarkers. The proposed model was evaluated against several classification algorithms, including Random Forest, Decision Tree, and Support Vector Classifier. Results demonstrated that the Optimized Cat Boost algorithm achieved the highest accuracy (0.93) and AUC (0.97), outperforming other models in precision, recall, and F1-score. This study highlights the potential of optimized machine learning techniques in improving the early detection of gastric cancer, offering a promising non-invasive diagnostic tool for clinical applications.

The release of pollutants, especially bisphenol A, the most important endocrine disruptor, from plastic recycling industries is critical. This study aims to examine the pollution load of bisphenol A (BPA) in the effluent of plastic recycling facilities located in the coastal provinces of the southern Caspian Sea to achieve a comprehensive understanding of their release into the environment. Physicochemical factors (Biochemical Oxygen Demand (BOD), Chemical Oxygen Demand (COD), and pH) were measured as influential factors. BPA measurement was conducted using gas chromatography-mass spectrometry after extraction. The average concentration of BPA, BOD, COD, and pH value in the effluent of recycling units without treatment was 0.24ppb, 2502.33mg/L, 8497.61mg/L and 7.44, and with treatment was 0.13ppb, 353.3mg/L, 1201.43mg/L and 6.92, respectively. The highest concentration of BPA is related to the effluent of the untreated polyethylene terephthalate (PET) unit in Gilan province (0.71ppb), which indicated the effect of the treatment in the reduction of pollutants. A positive correlation was observed between COD with BPA concentration (P < 0.05). The ecological risk of BPA for the effluent of PET units without treatment was 0.11ppb which was in moderate ecological risk and higher than the others.