http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/833 Volume 12, Number 2, December 2024 Sun, 26 Apr 2026 00:18:52 GMT 2026-04-26T00:18:52Z http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/865 Analyzing the Performance of Deep Learning Models for Detecting Hate Speech on Social Media Platforms Islam Arif, Md Ariful; Rahman, Md. Mahbubur; Rabiul Alam, Md. Golam; Akhtaruzzaman, M. Currently social media and online platforms have become a major source of cyberbullying and hate speech. It is currently affecting people and communities in harmful ways. Hate speech on social media is rising in Bangladesh and it is creating a need for effective tools to prevent and detect these incidents. This study introduces a deep learning model to mitigate this issue of identifying hate speech in text using three types of word embedding methods: Word2Vec, FastText, and BERT. The text data was labeled to mark hate speech and non-hate speech content. After that, these texts are preprocessed by removing punctuation and symbols to help improve model accuracy. Five deep learning models Bi-GRU-LSTM-CNN, Bi-LSTM, CNN, LSTM, and XGBoost were trained to classify the text as hate speech or non-hate speech. The study found that the LSTM model accomplished the highest accuracy at 95.66% with the Word2Vec embedding method, while CNN reached 87.70% with FastText embeddings. Word2Vec is effective for capturing word meanings in general text classification. FastText works well with rare words and languages that have complex word forms. These findings help advance effective hate speech detection techniques. It could promote more respectful and inclusive interactions on social media. This proposed deep-learning model can help stop cyberbullying and hate speech on social media. Sun, 01 Dec 2024 00:00:00 GMT http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/865 2024-12-01T00:00:00Z http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/864 RANS Flow Computation around Transonic RAE2822Airfoil with a New SST Turbulence Model Rahman, Md Mizanur; Zhang, Xueting; Hasan, K.; Chen, Sheng The solution to RANS (Reynolds-averaged Navier-Stokes)equations invokes a suitableframework for turbulence modelling. To account for turbulence and transitioneffects, a new SST (Shear Stress Transport)풌−흎turbulence model is coupled with RANS to simulate the transonic flow passing an RAE2822 air foil. Three sets of experimental data of the super-critical RAE2822 air foilare employed to validate the new SST (NSST) closure. Computations are conducted for a limited range of Reynolds numbers with variable angle of attack. The NSST model has been found to replicate satisfactory results for lift 푪푳and drag 푪푫coefficients as well as for skin-friction and pressure coefficient profiles under considerable shock-wave boundary layer (BL) interaction, although푪푫is challenging to be accurately predicted since the turbulence model requires to adequately resolve near-wall turbulence in the BL with varying pressure gradients. NSST predictions are compared with those of the widely-usedSST풌−흎model. Numerical outputsdemonstrate that the included NSST transition model plays no significant roles to appropriatelypredict푪푳and푪푫, indicating that the NSST performance is almost equivalent to that of the SST in the current analysis. Sun, 01 Dec 2024 00:00:00 GMT http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/864 2024-12-01T00:00:00Z http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/863 Numerical Simulation of a Single-tank Molten Salt Cell with Multifunctional Coupling Yang, Ying; Jin, Yingai; Sun, Yanwei Molten salt tanks are crucial in photovoltaic power plants, serving as the core of new energy-storage systems. Although suitable for small-area domestic heating, they are complex in structure and prone to significant heat loss. To address these issues, this studyproposes a novel single-tank molten salt system that combines monitoring, preheating, heat exchange, and storage functionalities. By incorporating a U-tube heat exchanger within the molten salt accumulator, the system achieves cost-effective heat storage and release. The effectiveness of the heat extraction method used with the U-tube profoundly affects the overall system performance. Through numerical simulations, this study examines the impact of different heat extraction techniques on the performance ofthe single-tank heat storage system, focusing on changes in the flow field within the molten salt during heat release. By modifying operational conditions, improvements in outlet temperature, heat release power, and heat utilization efficiency of the U-tube heat exchanger are demonstrated. This study explores the heat release process in a single tank of molten salt using 3D unsteady Computational Fluid Dynamics (CFD) simulations.Operation behaviour estimate results show thatvaried initial temperatures ofthe molten salt have distinct impacts on the thermal behavior of the system. Higher initial temperatures lead to a smaller temperature differential between the highest and lowest points in the tank during the same exothermic periods. And under conditions of constant inlet velocity, the exothermic power decreases as the duration of heat release increases. In scenarios with a constant inlet mass flow rate, the time required to reach the limit of exothermic power decreases as the mass flow rate increases.Throughout the exothermic process, the average heat flow density gradually declines. This decline is particularly notable in the first 10 minutes of the exothermic activity. As the process progresses, the average temperature through the heat transfer oil within the heat exchanger increases, which reduces the temperature differential between the hot and cold fluids, further decreasing the average heat flow density. Sun, 01 Dec 2024 00:00:00 GMT http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/863 2024-12-01T00:00:00Z http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/862 Bankline Assessment of Shibsa-Passur River by Satellite Imagery and Hydrodynamic Modeling using Delft3D Haque, Nusaiba Afsara; Azad, Anika; Amin, Md. Khairul; Hasan, G. M. Jahid The Sundarbans, a prolific mangrove wetland ecosystem part of the world's largest delta formed by the Ganges, Brahmaputra, and Meghna rivers, contribute immensely to coastal stability and protection. In deltaic mangroves such as the Sundarbans, recurrent erosion and accretion caused by the transportation of unconsolidated sediments by rivers, interact with flow velocity and bed shear to generate continuous morphological dynamics. This study analyzed the bankline migration patterns of the Shibsa and Passur rivers within the Bangladesh portion of the Sundarbans using satellite imagesand assessed their hydrodynamic behavior by developing a two-dimensional model using Delft3D. This study utilized Landsat and Sentinel images from 2009-2021 to identify critical erosion and deposition zones applyingthe DSAS tool of ArcGIS. A two-dimensional hydrodynamic model was then developed and calibrated to simulate flow velocity, bed shear stress, and water levels in these critical zones. The model was validated against limited available water level data. Six critical zones were identified, with four erosion-prone and two deposition-prone areas. The model results indicated increased velocities and bed shear stresses in the erosion zones relative to non-critical areas, while the deposition zones experienced reduced velocities and bed shear stresses. The study reveals that the high values of velocity and bed shear stressesare responsible for the morphological changesof erosion, thus emphasizing the significance of close monitoring with remotely sensed images. Sun, 01 Dec 2024 00:00:00 GMT http://dspace.mist.ac.bd:8080/xmlui/handle/123456789/862 2024-12-01T00:00:00Z