Browsing by Author "Mahmud, S. M. Hasan"

Now showing 1 - 13 of 13

A Novel Hybrid Approach for Classifying Osteosarcoma Using Deep Feature Extraction and Multilayer Perceptron
(MDPI, 2023-06-18) Aziz, Md. Tarek; Mahmud, S. M. Hasan; Elahe, Md. Fazla; Jahan, Hosney; Rahman, Md Habibur; Nandi, Dip; Smirani, Lassaad K.; Ahmed, Kawsar; Bui, Francis M.; Moni, Mohammad Ali
Osteosarcoma is the most common type of bone cancer that tends to occur in teenagers and young adults. Due to crowded context, inter-class similarity, inter-class variation, and noise in H&E-stained (hematoxylin and eosin stain) histology tissue, pathologists frequently face difficulty in osteosarcoma tumor classification. In this paper, we introduced a hybrid framework for improving the efficiency of three types of osteosarcoma tumor (nontumor, necrosis, and viable tumor) classification by merging different types of CNN-based architectures with a multilayer perceptron (MLP) algorithm on the WSI (whole slide images) dataset. We performed various kinds of preprocessing on the WSI images. Then, five pre-trained CNN models were trained with multiple parameter settings to extract insightful features via transfer learning, where convolution combined with pooling was utilized as a feature extractor. For feature selection, a decision tree-based RFE was designed to recursively eliminate less significant features to improve the model generalization performance for accurate prediction. Here, a decision tree was used as an estimator to select the different features. Finally, a modified MLP classifier was employed to classify binary and multiclass types of osteosarcoma under the five-fold CV to assess the robustness of our proposed hybrid model. Moreover, the feature selection criteria were analyzed to select the optimal one based on their execution time and accuracy. The proposed model achieved an accuracy of 95.2% for multiclass classification and 99.4% for binary classification. Experimental findings indicate that our proposed model significantly outperforms existing methods; therefore, this model could be applicable to support doctors in osteosarcoma diagnosis in clinics. In addition, our proposed model is integrated into a web application using the FastAPI web framework to provide a real-time prediction.
An Efficient Deep Learning Model to Infer User Demographic Information from Ratings
(Scopus, 2019-04-17) Mahmud, S. M. Hasan
Obtaining demographics of online users is of great significance to Internet service providers and advertisers. Most previous works used standard machine learning methods to infer user demographics from handcrafted features. This has two disadvantages. First, the handcrafted features are usually not robust and rely too much on expert experience. Second, these low-capacity models can neither model the complex non-linear relationship between users nor recognize interdependencies among items. To address these problems, we propose a DEep REtentive learning frameworK (DEREK) for demographic information prediction. Specifically, we introduce a heuristic data generation method that can alleviate data sparsity in order to use rating data more efficiently than the handcrafted features. Moreover, the retention blocks based on high-capacity deep neural networks are designed to extract a share representation from input rating data. Finally, the DEREK can simultaneously infer different demographic attributes through end-to-end multi-label learning architecture. The extensive experimental results on Movielens 100-k and Movielens 1-M data sets have demonstrated the superiority of the proposed DEREK compared with standard machine learning methods (logistic regression and SVM).
An LSTM network-based model with attention techniques for predicting linear T-cell epitopes of the hepatitis C virus
(Scopus, 2024) Hosen, Md. Faruk; Mahmud, S. M. Hasan; Goh, Kah Ong Michael; Uddin, Muhammad Shahin; Nandi, Dip; Shatabda, Swakkhar; Shoombuatong, Watshara
: Hepatitis C virus (HCV) infection remains a significant global health challenge, often resulting in severe longterm physical complexity and even death. Since its discovery, HCV has exhibited substantial genetic variability, complicating vaccine development. Although some therapeutic approach have shown efficacy against certain HCV genotypes, a universally effective vaccine is still lacking. Recent research suggests that the body’s cellular immune response, particularly T cell epitopes of HCV (TCE-HCVs), plays a vital role in fighting the virus. Therefore, the precise and rapid identification of TCE-HCVs is essential for chronic HCV infection. In this work, we proposed a novel TCE-HCVs prediction model AttLSTM, which combines attention mechanism and long shortterm memory (LSTM). Specifically, we employed four robust feature encoding techniques: One-Hot Encoding, Global Vectors (GloVe), fastText, and Word2Vec to encode protein sequences. Additionally, k-mer embedding was utilized to help the model identify significant subsequence fragments within the protein sequences. To optimize the model’s performance, irrelevant features are eliminated using the SHapley Additive exPlanations (SHAP) approach. The resulting optimal feature subset was then fed into the AttLSTM model to identify TCEHCVs. The attention mechanism in this model dynamically captures the pairwise correlations of each neighboring target pair within a sliding window, thereby enhancing the understanding of the local environment of target residues. Extensive experiments showed that AttLSTM outperformed conventional machine learning (ML) classifiers in predictive performance. Notably, in k-fold cross validation, AttLSTM achieved superior performance compared to existing methods with accuracy of 80.77 %, MCC of 0.632, and AUC of 0.891. This exceptional performance indicates that AttLSTM has a strong predictive capability for identifying TCE-HCVs. We anticipate that AttLSTM will expedite the rapid identification of promising TCE-HCVs, aiding in the development of diagnostic and immunotherapeutic treatments for HCV in the future.
Early-Stage Cervical Cancerous Cell Detection from Cervix Images Using YOLOv5
(Tech Science Press, 2023-01-01) Ontor, Md Zahid Hasan; Ali, Md Mamun; Ahmed, Kawsar; Bui, Francis M.; Al-Zahrani, Fahad Ahmed; Mahmud, S. M. Hasan; Azam, Sami
"Cervical Cancer (CC) is a rapidly growing disease among women throughout the world, especially in developed and developing countries. For this many women have died. Fortunately, it is curable if it can be diagnosed and detected at an early stage and taken proper treatment. But the high cost, awareness, highly equipped diagnosis environment, and availability of screening tests is a major barrier to participating in screening or clinical test diagnoses to detect CC at an early stage. To solve this issue, the study focuses on building a deep learning-based automated system to diagnose CC in the early stage using cervix cell images. The system is designed using the YOLOv5 (You Only Look Once Version 5) model, which is a deep learning method. To build the model, cervical cancer pap-smear test image datasets were collected from an open-source repository and these were labeled and preprocessed. Then the YOLOv5 models were applied to the labeled dataset to train the model. Four versions of the YOLOv5 model were applied in this study to find the best fit model for building the automated system to diagnose CC at an early stage. All of the model’s variations performed admirably. The model can effectively detect cervical cancerous cell, according to the findings of the experiments. In the medical field, our study will be quite useful. It can be a good option for radiologists and help them make the best selections possible."
Effects of Cognitive Ability, Trust and Time-saving
(Lecture Notes in Networks and Systems, Springer, 2018-08-31) Hossain, Rifat; Mahmud, S. M. Hasan; Hossin, Md. Altab; Bhuiyan, Touhid; Hua, Yin Xiao
Prior research was perused the technology usage on limited factors; our aim is to investigate factor to adumbrate e-ticketing systems accuracy with further improvement and successive usage in the context of Bangladesh. To do so, we designed a research model with the integration of Technology Acceptance Model (TAM), Theory Planned Behaviour (TPB), trust theory, time-saving theory and cognitive framework. Survey data were amassed from the different profession of consumers who are using e-ticketing system; 145 valid data were tested the research model to find structural relationship and hypothesis result by using Partial Least Squares (PLS). Cognitive ability has a significant impact on time-saving and trust to purchase the ticket on Internet, while a wispy impact has cited between cognitive ability and perceived usefulness. In addition, trust has positively influenced on attitude towards of buying an online ticket, and time-saving has a positive relationship with TAM. Moreover, the study assists e-ticketing authorities and individual consumer to understand which factor gradually affects the progressive usage of e-ticketing.
Factors Impacting Short-Term Load Forecasting of Charging Station to Electric Vehicle
(MDPI Publications, 2023-12-23) Elahe, Md Fazla; Kabir, Md Alamgir; Mahmud, S. M. Hasan; Azim, Riasat
The rapid growth of electric vehicles (EVs) is likely to endanger the current power system. Forecasting the demand for charging stations is one of the critical issues while mitigating challenges caused by the increased penetration of EVs. Uncovering load-affecting features of the charging station can be beneficial for improving forecasting accuracy. Existing studies mostly forecast electricity demand of charging stations based on load profiling. It is difficult for public EV charging stations to obtain features for load profiling. This paper examines the power demand of two workplace charging stations to address the above-mentioned issue. Eight different types of load-affecting features are discussed in this study without compromising user privacy. We found that the workplace EV charging station exhibits opposite characteristics to the public EV charging station for some factors. Later, the features are used to design the forecasting model. The average accuracy improvement with these features is 42.73% in terms of RMSE. Moreover, the experiments found that summer days are more predictable than winter days. Finally, a state-of-the-art interpretable machine learning technique has been used to identify top contributing features. As the study is conducted on a publicly available dataset and analyzes the root cause of demand change, it can be used as baseline for future research.
IDTi-CSsmoteB
(IEEE Access, 2019-04-11) Mahmud, S. M. Hasan; Chen, Wenyu; Jahan, Hosney; Liu, Yongsheng; Sujan, Nasir Islam; Ahmed, Saeed
Identifying interaction between drug and protein is a crucial challenge in drug discovery, which can lead the researchers to develop novel drug compounds or new target proteins for the existing drugs. The determination of drug-target interactions (DTIs) is an extremely time-consuming, costly, and tedious task with wet-lab experiments. To date, multiple computational techniques have been presented to simplify the drug discovery process, but a huge number of interactions are still undiscovered. Furthermore, a class imbalance is a critical challenge regarding this experiment which can significantly degrade the classification accuracy that has not been effectively addressed yet. In this paper, we proposed a novel high-throughput computational model, called iDTi-CSsmoteB, for identification of DTIs based on drug chemical structures and protein sequences. More specifically, the protein sequence is extracted through position-specific scoring matrix (PSSM)-Bigram, amphiphilic pseudo amino acid composition (AM-PseAAC) and dipeptide PseAAC descriptors which represents evolutionary and sequence information. The drug chemical structure is represented as a molecular substructure fingerprint (MSF) which describes the existence of the functional fragments or groups. Finally, we used the over-sampling SMOTE technique to overcome the imbalance issue of the datasets and applied XGBoost algorithm as a classifier to predict DTIs. To evaluate the performance of iDTi-CSsmoteB, several experiments have been conducted on four benchmark datasets, namely, enzyme, ion channel, GPCR, and nuclear receptor based on fivefold cross validation. The experimental analysis exhibits that our model outperforms similar methods in terms of area under the ROC (auROC) curve. In addition, our achieved results indicate the effectiveness of the feature extraction techniques, balancing methods, and classifier for predicting the DTIs which can provide substance for new drug development. iDTi-CSsmoteB webserver is available online at http://idticssmoteb-uestc.me/
Machine Learning Approach for Software Defect Prediction
(Lecture Notes in Electrical Engineering, Springer, 2020-03-24) Hossen, Md Anwar; Islam, Md. Shariful; Yusof, Nurhafizah Abu Talip; Rahman, Md. Sakib; Siddika, Fatema; Rahman, Mostafijur; Khatun, Sabira; Karim, Mohamad Shaiful Abdul; Mahmud, S. M. Hasan
The software has turn into an imperious part of human’s life. In the recent computing era, many large-scale complex network systems and millions of modern technological devices produce a huge amount of data every second. Among these data, the amount of imbalanced data is relatively excessive. The machine learning model is miss leaded by these imbalanced data. Software Defect Prediction (SDP) is a standout amongst the most helping exercises during the testing phase. The estimated cost of finding and fixing defects is approximately billions of pounds per year. To reduce this problem, software defect prediction has come forth but need fine tuning to have expected efficiency. In this chapter, we have proposed a new model based on machine learning approach to predict software defect and identify the key factors that may help the software engineer to identify the most defect-prone part of the system. The proposed model works as follows. First, need to remove highly correlated features and turn all the feature in the same scale using the scaling feature approach. Second, we have used Synthetic Minority Over-Sampling Technique (SMOTE), Adaptive Synthetic (ADASYN) and Hybrid sampling method to balance highly imbalanced datasets. Third, Random Forest Importance and Chi-square algorithms are chosen to find out the factors which have high effect on software defect. Cross validation is used to remove overriding problem. Scikit-learn library is used for machine learning algorithms. Pandas library is used for data processing. Matplotlib, and PyPlot are used for graph and data visualization respectively. The hybrid sampling method and Random Forest (RF) algorithms achieved the highest prediction accuracy about 93.26% by showing its superiority.
Publishing CSV Data as Linked Data on the Web
(Scopus, 2020) Mahmud, S. M. Hasan; Hossin, Md. Altab; Hasan, Md. Rezwan; Jahan, Hosney; Noori, Sheak Rashed Haider; Ahmed, Md. Razu
The majority of datasets on Open Government Data (OGD) portals are stored in comma-separated values (CSV) file. Publishing CSV data as a Linked Open Data (LOD) on the Web is an active field of research. However, there are very few effective applications have been developed with this purpose. Linked Data refer many ways for connecting and publishing structured data to data consumers, but available datasets are in CSV format. Therefore, publishing the CSV model on the webpage, it is needed to change CSV in RDF file format. Many methods and tools have been proposed for data mapping and publishing, however, most of them are not followed by the W3C recommendations rules. The contribution and goal of this paper are to develop a Semantic approach that can effectively convert CSV data into RDF data with rich semantics and release RDF data on the web using LOD principles. We utilize Semantic Web resources and W3C recommendation rules in automatic data publishing method, which enables distributed system for scalability. We apply the proposed method to existing CSVW Implementation Report-W3C and U.S Government’s application (data.gov). Our experimental results indicate that the proposed approach successfully converts CSV to RDF data and publish those RDF as LOD on the Web, with adequate performance on any sized datasets.
Systems Biology Models To Identify the Influence of SARS-CoV-2 Infections to the Progression of Human Autoimmune Diseases
(Daffodil International University, 2022-08-02) Al-Mustanjid, Md.; Mahmud, S. M. Hasan; Akter, Farzana; Rahman, Md Shazzadur; Hossen, Md Sajid; Rahman, Md Habibur; Moni, Mohammad Ali
Severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) has been circulating since 2019, and its global dominance is rising. Evidences suggest the respiratory illness SARS-CoV-2 has a sensitive affect on causing organ damage and other complications to the patients with autoimmune diseases (AD), posing a significant risk factor. The genetic interrelationships and molecular appearances between SARS-CoV-2 and AD are yet unknown. We carried out the transcriptomic analytical framework to delve into the SARS-CoV-2 impacts on AD progression. We analyzed both gene expression microarray and RNA-Seq datasets from SARS-CoV-2 and AD affected tissues. With neighborhood-based benchmarks and multilevel network topology, we obtained dysfunctional signaling and ontological pathways, gene disease (diseasesome) association network and protein-protein interaction network (PPIN), uncovered essential shared infection recurrence connectivities with biological insights underlying between SARS-CoV-2 and AD. We found a total of 77, 21, 9, 54 common DEGs for SARS-CoV-2 and inflammatory bowel disorder (IBD), SARS-CoV-2 and rheumatoid arthritis (RA), SARS-CoV-2 and systemic lupus erythematosus (SLE) and SARS-CoV-2 and type 1 diabetes (T1D). The enclosure of these common DEGs with bimolecular networks revealed 10 hub proteins (FYN, VEGFA, CTNNB1, KDR, STAT1, B2M, CD3G, ITGAV, TGFB3). Drugs such as amlodipine besylate, vorinostat, methylprednisolone, and disulfiram have been identified as a common ground between SARS-CoV-2 and AD from drug repurposing investigation which will stimulate the optimal selection of medications in the battle against this ongoing pandemic triggered by COVID-19.
The comparative analysis of online shopping information platform's security based on customer satisfaction
(IEEE Xplore, 2017-10-19) Mahmud, S. M. Hasan; Kabir, Md Alamgir; Salem, Omar A. M.; Fernand, Kazihise Ntikurako Guy
With the massive expansion of Internet and e-commerce technology, Internet platform is providing a lot of services and advantages for online business, especially for online shopping portal. As a result of the Internet, online shopping has expanded in businesses more effectively and online services are collaborating with customers and other associations. For improving online shopping information platform's security, customer satisfaction is one of the most fundamental factors. This paper constructs the measurement of four-dimensional models that are appropriate for measuring customer satisfaction of online shopping information platform's security. This paper also conducts the factor and multiple regression analysis to verify the measurement model. By using questionnaires survey and analysis from the groups of Hohai University students, this research provides suggestions for the development of large online shopping portal and will help to improve customer satisfaction on the security of the online shopping information platforms. Full Text Link: http://doi.org/10.1109/ICCSNT.2016.8070139
Training Data Selection Using Ensemble Dataset Approach for Software Defect Prediction
(Scopus, 2020) Sohan, Md Fahimuzzman; Kabir, Md Alamgir; Rahman, Mostafijur; Mahmud, S. M. Hasan; Bhuiyan, Touhid
Cross-project defect prediction (CPDP) is using due to the limitation of within project defect prediction (WPDP) in Software Defect Prediction (SDP) research. CPDP aims to train one project data to predict another project using the machine learning technique. The source and target projects are different in the CPDP setting, because of various structured source-target projects, sometimes it may not be a perfect combination. This study represents a categorical data set ensemble technique, where multiple data sets have been aggregated for source data instead of using a single data set. The method has been evaluated on nine data sets, taken from the publicly accessible repository with two performance indicators. The results of this data set ensemble approach show the improvement of the prediction performance over 65% combinations compared with traditional CPDP models. The results also show that same categories (homogeneous) train-test data set pairs give high performance; otherwise, the prediction performances of different category data sets are mostly collapsed. Therefore, the proposed scheme is recommended as an alternative to predict defects that can improve the prediction of most of the cases compared with traditional cross-project SDP models.
YOLO_CC: Deep Learning based Approach for Early Stage Detection of Cervical Cancer from Cervix Images Using YOLOv5s Model
(Daffodil International University, 2022-08-26) Ontor, Md Zahid Hasan; Ali, Md Mamun; Ahmed, Kawsar; Bui, Francis M.; Al-Zahrani, Fahad Ahmed; Mahmud, S. M. Hasan; Azam, Sami
Cervical Cancer (CC) is the fourth major cancer, which is responsible for a large number of deaths among women. Early stage detection of the cancer is the most effective solutions for decreasing the mortality rate. The lack of awareness, and costly clinical diagnosis are the major barrier for women to participate in clinical screening test to detect CC at an early stage. To address the issue, the study aims to find a deep learning based intelligent system to detect CC in the early stage using real time images. To build the model, cervical cancer pap-smear test image datasets were gathered and these were labeled and preprocessed. Then the YOLOv5 model was employed on the labeled dataset to train the model. Three latest versions of YOLOv5 model were applied in this study to find the most efficient model for building the intelligent system to detect CC at an early stage. All of the applied models provided satisfactory performance. Among all the applied models, YOLOv5s outperformed with 0.8279 precision and 0.8265 recall value. The performance of the study indicates that the proposed model highly potential to diagnose CC using real time images in early stage. In the medical field, the proposed will be quite useful for clinicians, and medical professionals.(CC) is the fourth major cancer, which is responsible for a large number of deaths among women. Early stage detection of the cancer is the most effective solutions for decreasing the mortality rate. The lack of awareness, and costly clinical diagnosis are the major barrier for women to participate in clinical screening test to detect CC at an early stage. To address the issue, the study aims to find a deep learning based intelligent system to detect CC in the early stage using real time images. To build the model, cervical cancer pap-smear test image datasets were gathered and these were labeled and preprocessed. Then the YOLOv5 model was employed on the labeled dataset to train the model. Three latest versions of YOLOv5 model were applied in this study to find the most efficient model for building the intelligent system to detect CC at an early stage. All of the applied models provided satisfactory performance. Among all the applied models, YOLOv5s outperformed with 0.8279 precision and 0.8265 recall value. The performance of the study indicates that the proposed model highly potential to diagnose CC using real time images in early stage. In the medical field, the proposed will be quite useful for clinicians, and medical professionals.