Browsing by Author "Islam, Mirajul"

Now showing 1 - 12 of 12

A Comprehensive Study of DCNN Algorithms-based Transfer Learning for Human Eye Cataract Detection
(IJACSA, 2023-06-01) Jidan, Omar Jilani; Paul, Susmoy; Roy, Anirban; Khushbu, Sharun Akter; Islam, Mirajul; Badhon, S.M. Saiful Islam
This study presents a comparative analysis of different deep convolutional neural network (DCNN) architectures, including VGG19, NASNet, ResNet50, and MobileNetV2, with and without data augmentation, for the automatic detection of cataracts in fundus images. Utilizing hybrid architecture models, namely ResNet50-NASNet and ResNet50+MobileNetV2, which combine two state-of-the-art DCNNs, this research demonstrates their superior performance. Specifically, MobileNetV2 and the combined ResNet50+MobileNetV2 outperform other models, achieving an impressive accuracy of 99.00%. By emphasizing the efficacy of diverse datasets and pre-processing techniques, as well as the potential of pretrained DCNN models, this study contributes to accurate cataract diagnosis. Furthermore, the proposed system has the potential to reduce reliance on ophthalmologists, decrease the cost of eye check-ups, and improve accessibility to eye care for a wider population. These findings showcase the successful application of deep learning and image processing techniques in the early detection and treatment of various medical conditions, including cataracts, addressing the needs of individuals with diminished vision through ocular images and innovative hybrid architectures.
An LSTM network-based model with attention techniques for predicting linear T-cell epitopes of the hepatitis C virus
(Scopus, 2024) Ahmed, Md. Kawsar; Nahin, Kamal Hossain; Ahammed, Md. Sharif; Haque, Md. Ashraful; Singh, Narinderjit Singh Sawaran; Ananta, Redwan Al Mahmud Asad; Nirob, Jamal Hossain; Islam, Mirajul; Paul, Liton Chandra
In this research, we explain comprehensive industrial and innovation results on using an artificial neural network (ANN) method to improve the performance of microstrip patch antennas for 5G, indoor-outdoor, and Ku band uses. To determine if an antenna is appropriate, this article discusses multiple methods, one of which is to do a simulation using validating software like high frequency structure simulator (HFSS) and Altair Feko. Based on the Rogers RT 5880 substrate, the antenna is constructed. There is a loss tangent of 0.0009 and its dimensions are 17.1053 mm in length and 16 mm in width. Its dielectric constant is 2.2. Despite its small size, it boasts an impressive maximum efficiency of almost 90% and a gain of approximately 8 dB. As an indicator of ANN model performance, we may look at the R-squared value (99%), the mean square error (MSE), which is approximately 0.0015, and the confidence interval (99%). The ANN models are the most accurate and have the lowest error rate when it comes to predicting efficiency and gain. The suggested antenna is a promising contender for the targeted Ku band, indoor/outdoor, and 5G uses, as verified by the clustering of computer simulation technology (CST), HFSS, and Altair Feko simulated results with the measured and predicted outcomes of ANN approach
BTSD: A Curated Transformation of Sentence Dataset for Text Classification in Bangla Language
(Elsevier, 2023-07-24) Das, Rajesh Kumar; Islam, Mirajul; Khushbu, Sharun Akter
The Bangla Transformation of Sentence Classification dataset addresses the resource gap in natural language processing (NLP) for the Bangla language by providing a curated resource for Bangla sentence classification. With 3,793 annotated sentences, the dataset focuses on categorizing Bangla sentences into Simple, Complex, and Compound classes. It serves as a benchmark for evaluating NLP models on Bangla sentence classification, promoting linguistic diversity and inclusive language models. Collected from publicly accessible Facebook pages, the dataset ensures balanced representation across the categories. Preprocessing steps, including anonymization and duplicate removal, were applied. Three native Bangla speakers independently assessed the Transformation of Sentence labels, enhancing the dataset's reliability. The dataset empowers researchers, practitioners, and developers to build accurate and robust NLP models tailored to the Bangla language. It offers insights into Bangla syntax and structure, benefiting linguistic research. The dataset can be used to train models, uncover patterns in Bangla language usage, and develop effective NLP applications across domains.
DCNN Based Disease Prediction of Lychee Tree
(Springer, 2023-04-17) Islam, Saiful; Akter, Shornaly; Islam, Mirajul; Rahman, Md. Arifur
Tree disease classification is needed to determine the affected leaves as it controls the economic importance of the trees and their products and decreases their eco-friendly eminence. The lychee tree is affected by some of the diseases named Leaf Necrosis, Stem Canker and leaf spots. Therefore, classifying the Lychee tree is essential to find the good and affected leaves. Our economic growth will be very high if we can adequately do the Lychee tree classification. In this paper, we tried to do a Lychee tree disease classification to make things easier for the farmers as they cannot correctly distinguish the good and bad leaves in an earlier stage. We have created a new data set for training the architectures. We have collected about 1400 images with three categories of pre-harvest diseases “Leaf Necrosis”, “Leaf Spots”, and “Stem Canker”. There are 1400 images in total, and out of those, 80% of the data is for training and 20% is for testing, this dataset has fresh and affected leaves and stems. For Lychee tree disease classification, we have chosen pre-trained CNN and Transfer Learning based approach to classify the layer of the 2D image by layer. This method can classify images efficiently from the images of disease leaves and stems. It will address disease from the images of the leaves and trees and determine specific preharvest diseases.
Deep Learning Based Classification System for Recognizing Local Spinach
(Springer, 2022-01-03) Islam, Mirajul; Ria, Nushrat Jahan; Ani, Jannatul Ferdous; Masum, Abu Kaisar Mohammad; Abujar, Sheikh; Hossain, Syed Akhter
A deep learning model gives an incredible result for image processing by studying from the trained dataset. Spinach is a leaf vegetable that contains vitamins and nutrients. In our research, a Deep learning method has been used that can automatically identify spinach and this method has a dataset of a total of five species of spinach that contains 3785 images. Four Convolutional Neural Network (CNN) models were used to classify our spinach. These models give more accurate results for image classification. Before applying these models there is some preprocessing of the image data. For the preprocessing of data, some methods need to happen. Those are RGB conversion, filtering, resize and rescaling, and categorization. After applying these methods image data are preprocessed and ready to be used in the classifier algorithms. The accuracy of these classifiers is in between 98.68 and 99.79%. Among those models, VGG16 achieved the highest accuracy of 99.79%.
Diabetes Among Adults in Bangladesh
(Daffodil International University, 2022-08-05) Chowdhury, Muhammad Abdul Baker; Islam, Mirajul; Rahman, Jakia; Uddin, Md Jamal; Haque, Md. Rabiul
Objective/research question To investigate the change in the prevalence and risk factors of diabetes among adults in Bangladesh between 2011 and 2018. Design The study used two waves of nationally representative cross-sectional data extracted from the Bangladesh Demographic and Health Surveys in 2011 and 2017–2018. Setting Bangladesh. Participants 14 376 adults aged ≥35 years. Primary outcome Diabetes mellitus (type 2 diabetes). Results From 2011 to 2018, the diabetes prevalence among adults aged ≥35 years increased from 10.95% (880) to 13.75% (922) (p<0.001), with the largest-relative increase (90%) among obese individuals. Multivariable logistic regression analysis identified age and body mass index (BMI) were the key risk factors for diabetes. Adults who were overweight or obese were 1.54 times (adjusted OR (AOR): 1.54, 95% CI: 1.20 to 1.97) more likely to develop diabetes than normal-weight individuals in 2011, and 1.22 times (AOR: 1.22, 95% CI: 1.00 to 1.50) and 1.44 times (AOR: 1.44, 95% CI: 1.13 to 1.84) more prone to develop diabetes in 2018. Other significant risk factors for diabetes were marital status, education, geographical region, wealth index and hypertension status in both survey years. Conclusion A high prevalence of diabetes was observed and it has been steadily increasing over time. To enhance diabetes detection and prevention among adults in Bangladesh, population-level interventions focusing on health education, including a healthy diet and lifestyle, are required.
Diabetes Prediction at Early Stage Using Machine Learning
(©Daffodil International University, 2022-01-13) Chowdhury, Md. Zaman; Islam, Mirajul; Tuly, Israt Jahan
The categorization of medical datasets using machine learning has piqued the interest of the academic community in recent years, despite the fact that it is a difficult undertaking. The use of a large number of machine learning algorithms to a collection of data aids in the completion of the processes. Many studies have been conducted in the past to predict disease using machine learning. However, there are several opportunities for improvement. The goal of this study is to show how pre-processing approaches, as well as traditional and ensemble classifiers, may be used to compare different machine learning-based models for diabetes prediction. The pre-processing procedures for processing the dataset include encoding categorical data, imputing missing values, handling imbalanced data, and scaling features are taken place in this exploration. Five classification techniques, including Support Vector Machine (SVM), Naive Bayes (NB), Logistic Regression (LR), Decision Tree (DT), and Extra Tree (ET), are used to classify the dataset using a 10-fold crossvalidation technique, as well as hyperparameter tuning in each classifier to assign the best parameters. Ensemble methods are used to improve the performance of traditional algorithms and prevent them from being over fitted and biased. The experimental study indicates diabetes predictions with a higher degree of accuracy, as well as evaluated the findings of other current studies, with 98.44% accuracy being the best.
Dual Band Antenna Design and Prediction of Resonance Frequency Using Machine Learning Approaches
(Scopus, 22-10-18) Haque, Md. Ashraful; Sarker, Nayan; Singh, Narinderjit Singh Sawaran; Rahman, Md Afzalur; Hasan, Md. Nahid; Islam, Mirajul; Zakariya, Mohd Azman; Paul, Liton Chandra; Sharker, Adiba Haque; Abro, Ghulam E. Mustafa; Hannan, Md; Pk, Ripon
An inset fed-microstrip patch antenna (MPA) with a partial ground structure is constructed and evaluated in this paper. This article covers how to evaluate the performance of the designed antenna by using a combination of simulation, measurement, creation of the RLC equivalent circuit model, and the implementation of machine learning approaches. The MPA’s measured frequency range is 7.9–14.6 GHz, while its simulated frequency range is 8.35–14.25 GHz in CST microwave studio (CST MWS) 2018. The measured and simulated bandwidths are 6.7 GHz and 5.9 GHz, respectively. The antenna substrate is composed of FR-4 Epoxy, which has a dielectric constant of 4.4 and a loss tangent of 0.02. The equivalent model of the proposed MPA is developed by using an advanced design system (ADS) to compare the resonance frequencies obtained by using CST. In addition, the measured return loss of the prototype is compared with the simulated return loss observed by using CST and ADS. At the end, 86 data samples are gathered through the simulation by using CST MWS, and seven machine learning (ML) approaches, such as convolutional neural network (CNN), linear regression (LR), random forest regression (RFR), decision tree regression (DTR), lasso regression, ridge regression, and extreme gradient boosting (XGB) regression, are applied to estimate the resonant frequency of the patch antenna. The performance of the seven ML models is evaluated based on mean square error (MSE), mean absolute error (MAE), root mean square error (RMSE), and variance score. Among the seven ML models, the prediction result of DTR (MSE = 0.71%, MAE = 5.63%, RMSE = 8.42%, and var score = 99.68%) is superior to other ML models. In conclusion, the proposed antenna is a strong contender for operating at the entire X-band and lower portion of the Ku-band frequencies, as evidenced by the simulation results through CST and ADS, it measured and predicted results using machine learning approaches.
Multihead Text Mining from COVID-19 Feedback Using Machine Learning, Deep Learning, and Hybrid Deep Learning Approaches
(2024-08-24) Kobra, Khadijatul; Sammi, Samrina Sarkar; Rahman, Naimur; Khushbu, Sharun Akter; Islam, Mirajul
This study examines the impact of the COVID-19 epidemic on students in Bangladesh through text classification using various machine learning (ML) algorithms and deep learning (DL) models. The pandemic led to emergency crisis protocols in the country, including self-quarantine and the closure of educational and governmental institutions, resulting in significant negative impacts on individuals’ physical and mental health, including anxiety, sadness, and terror. To better understand the psychological effects of the epidemic, the authors collected survey data from 400 students in various divisions of Bangladesh using self-administered questionnaires through Google Forms. Preprocessing techniques such as tokenization, filtering, and n-gram modeling were used in the analysis. The study deployed eight different ML algorithms and DL models, including LSTM, BiLSTM, and CNN, to classify the effects on students’ academic, mental, and social lives. The results show that the ML classifier algorithms were highly effective, achieving accuracies of 95.00%, 93.75%, and 95.00% for academic, mental, and social life impact, respectively. Furthermore, hybrid DL models, such as CNN-LSTM and CNN-BiLSTM, produced good scores in predicting the impacts on students’ lives. Overall, this study provides valuable insights into the impacts of the COVID-19 epidemic on students’ academic, mental, and social well-being in Bangladesh.
Quasi-Yagi Antenna Design for LTE Applications and Prediction of Gain and Directivity Using Machine Learning Approaches
(Elsevier, 2023-09-01) Haque, Md. Ashraful; Zakariya, M.A.; Al-Bawri, Samir Salem; Yusoff, Zubaida; Islam, Mirajul; Saha, Dipon; Abdulkawi, Wazie M.; Rahman, Md Afzalur; Paul, Liton Chandra
In recent years, improvements in wireless communication have led to the development of microstrip or patch antennas. The article discusses using simulation, measurement, an RLC equivalent circuit model, and machine learning to assess antenna performance. The antenna's dimensions are 1.01 with respect to the lowest operating frequency, the maximum achieved gain is 6.76 dB, the maximum directivity is 8.21 dBi, and the maximum efficiency is 83.05%. The prototype's measured return loss is compared to CST and ADS simulations. The prediction of gain and directivity of the antenna is determined using a different supervised regression machine learning (ML) method. The performance of ML models is measured by the variance score, R square, mean square error (MSE), mean absolute error (MAE), root mean square error (RMSE), and mean squared logarithmic error (MSLE), etc. With errors of less than unity and an accuracy of roughly 98%, Ridge regression gain prediction outperforms the other seven ML models. Gaussian process regression is the best method for predicting directivity. Finally, modeling results from CST and ADS, as well as measured and anticipated results from machine learning, reveal that the suggested antenna is a good candidate for LTE.
Sentiment Analysis in Multilingual Context: Comparative Analysis of Machine Learning and Hybrid Deep Learning Models
(Elsevier, 2023-09-19) Das, Rajesh Kumar; Islam, Mirajul; Hasan, Md Mahmudul; Razia, Sultana; Hassan, Mocksidul; Khushbu, Sharun Akter
This research paper investigates the efficacy of various machine learning models, including deep learning and hybrid models, for text classification in the English and Bangla languages. The study focuses on sentiment analysis of comments from a popular Bengali e-commerce site, "DARAZ," which comprises both Bangla and translated English reviews. The primary objective of this study is to conduct a comparative analysis of various models, evaluating their efficacy in the domain of sentiment analysis. The research methodology includes implementing seven machine learning models and deep learning models, such as Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Convolutional 1D (Conv1D), and a combined Conv1D-LSTM. Preprocessing techniques are applied to a modified text set to enhance model accuracy. The major conclusion of the study is that Support Vector Machine (SVM) models exhibit superior performance compared to other models, achieving an accuracy of 82.56% for English text sentiment analysis and 86.43% for Bangla text sentiment analysis using the porter stemming algorithm. Additionally, the Bi-LSTM Based Model demonstrates the best performance among the deep learning models, achieving an accuracy of 78.10% for English text and 83.72% for Bangla text using porter stemming. This study signifies significant progress in natural language processing research, particularly for Bangla, by enhancing improved text classification models and methodologies. The results of this research make a significant contribution to the field of sentiment analysis and offer valuable insights for future research and practical applications.
Urban insights into mental health:
(Daffodil International University, 2023-11) Shahjahan, Md; Islam, Md. Mazharul; Islam, Mirajul; Das, Kumer Pial
This study investigates the prevalence of depression and anxiety symptoms among urban adults in Bangladesh and explores the associated socio-demographic factors. An online cross-sectional survey was conducted from March to May 2021 among the adults living in Dhaka City, Bangladesh, using a standard questionnaire designed using Google Forms and sent via a unique uniform resource locator (URL). The Depression Anxiety Stress Scale 21 (DASS-21) was employed to assess symptoms of mental health issues. The survey included responses from 993 participants. The results indicate that about 40% of the urban adults had moderate-severe anxiety symptoms, while about 46% had depression symptoms. Urban adults’ level of education, gender, occupation, place of birth, and asset quintile were found to have a significant correlation with increased anxiety and depression. The findings stressed the need for developing effective mental health services and educational programs for urban people to meet their mental health needs and prevention mechanisms. There is an urgent need for interventions that specifically target mental health issues.