Browsing by Author "Noori, Sheak Rashed Haider"

Now showing 1 - 20 of 34

A Deep Learning Approach to Predict Chronic Kidney Disease in Human
(Scopus, 2021) Arafat, Faisal; Khan, Thaharim; Bapon, Atanu Das; Khan, Md. Ibrahim; Noori, Sheak Rashed Haider
Renal turmoil otherwise called Chronic Kidney Disease (CKD) has been a very important field of study for a long while now. Diagnosis of CKD requires a lot of tests and it's not a straightforward or easy process. Recent advancements in machine learning (ML) based disease classification have attracted researchers to investigate various health data. The aim of this article is to automate the detection process of CKD using clinical data by employing a deep learning (DL) model. Moreover, this study intends to achieve a robust and feasible model to detect the CKD with comprehensive clinical accuracy. Initially, preprocessing and feature engineering tasks have been performed on a dataset having 400 instances and 23 attributes. Finally, the dataset was fed to the deep learning model to classify the diagnosis of CKD. This research has obtained a higher accuracy (99%) than other recently utilized methods in CKD diagnosis by employing the deep learning model.
A Novel Data and Model Centric Artificial Intelligence Based Approach in Developing High-Performance Named Entity Recognition for Bengali Language
(PLOS ONE, 2023-09-22) Lima, Khadija Akter; Hasib, Khan Md; Azam, Sami; Karim, Asif; Montaha, Sidratul; Noori, Sheak Rashed Haider; Jonkman, Mirjam
Named Entity Recognition (NER) plays a significant role in enhancing the performance of all types of domain specific applications in Natural Language Processing (NLP). According to the type of application, the goal of NER is to identify target entities based on the context of other existing entities in a sentence. Numerous architectures have demonstrated good performance for high-resource languages such as English and Chinese NER. However, currently existing NER models for Bengali could not achieve reliable accuracy due to morphological richness of Bengali and limited availability of resources. This work integrates both Data and Model Centric AI concepts to achieve a state-of-the-art performance. A unique dataset was created for this study demonstrating the impact of a good quality dataset on accuracy. We proposed a method for developing a high quality NER dataset for any language. We have used our dataset to evaluate the performance of various Deep Learning models. A hybrid model performed with the exact match F1 score of 87.50%, partial match F1 score of 92.31%, and micro F1 score of 98.32%. Our proposed model reduces the need for feature engineering and utilizes minimal resources.
A Study Review of Neural Audio Speech Transposition over Language Processing
(IEEE, 2023-04-17) Khushbu, Sharun Akter; Ajmain, Moshfiqur Rahman; Rahman, Mahafozur; Noori, Sheak Rashed Haider
Natural Language processing is the advancement of Artificial Intelligence in the modern technological era. Machine Translation this is the vast majority of languages transformation among human languages and computer interaction. NLP domain creates a sequential analysis of the path where the neural network basement is mathematically and theoretically strong enough. In unwritten language aim to multimodal language transformation. According to the spoken language there are several aspects of prosperity. Thereby coming up with the development of linguistic CNN models for all manpower who spoke in their mother tongue. Therefore, concern with english speech language processing advancement has a great impact on language transformation. In a sense other languages can be placed by the speech to language transformation computational period emerged on severally made corpus languages. Due to this implementation of model refer probabilistic model. Consequently, this is a study of a benchmark of recent happenings in unwritten language to language modeling in which summarization or transformation will be faster. With concern current research work that has described how performing best in CNN model and attention model, described statistical deployment. Finally, this paper is capable of inflicting the solution. Furthermore, studies have discussed the impact of result, observation, challenges and limitation with the respect of the solution. The challenge is voice identification without noise is challenging.
A Systematic Review of Graph Neural Network in Healthcare-Based Applications: Recent Advances, Trends, and Future Directions
(Scopus, 2024-01-16) Saha, Arpa; Hasan, Md. Zahid; Noori, Sheak Rashed Haider; Moustafa, Ahmed
Graph neural network (GNN) is a formidable deep learning framework that enables the analysis and modeling of intricate relationships present in data structured as graphs. In recent years, a burgeoning interest has arisen in exploiting the latent capabilities of GNN for healthcare-based applications, capitalizing on their aptitude for modeling complex relationships and unearthing profound insights from graph-structured data. However, to the best of our knowledge, no study has systemically reviewed the GNN studies conducted in the healthcare domain. This study has furnished an all-encompassing and erudite overview of the prevailing cutting-edge research on GNN in healthcare. Through analysis and assimilation of studies, current research trends, recurrent challenges, and promising future opportunities in GNN for healthcare applications have been identified. China emerged as the leading country to conduct GNN-based studies in the healthcare domain, followed by the USA, UK, and Turkey. Among various aspects of healthcare, disease prediction and drug discovery emerge as the most prominent areas of focus for GNN application, indicating the potential of GNN for advancing diagnostic and therapeutic approaches. This study proposed research questions regarding diverse aspects of GNN in the healthcare domain and addressed them through an in-depth analysis. This study can provide practitioners and researchers with profound insights into the current landscape of GNN applications in healthcare and can guide healthcare institutes, researchers, and governments by demonstrating the ways in which GNN can contribute to the development of effective and efficient healthcare systems
Analysis of Bangla Transformation of Sentences Using Machine Learning
(Springer, 2023-04-17) Das, Rajesh Kumar; Sammi, Samrina Sarkar; Kobra, Khadijatul; Ajmain, Moshfiqur Rahman; khushbu, Sharun Akter; Noori, Sheak Rashed Haider
In many languages, various language processing tools have been developed. The work of the Bengali NLP is getting richer day by day. Sentence pattern recognition in Bangla is a subject of attention. Additionally, our motivation was to work on implementing this pattern recognition concept into user-friendly applications. So, we generated an approach where a sentence (sorol, jotil and jougik) can be correctly identified. Our model accepts a Bangla sentence as input, determines the sentence construction type, and outputs the sentence type. The most popular and well-known six supervised machine learning algorithms were used to classify three types of sentence formation: Sorol Bakko (simple sentence), Jotil Bakko (complex sentence) and Jougik Bakko(compound sentence). We trained and tested our dataset, which contains 2727 numbers of data from various sources. We analyzed our dataset and got accuracy, precision, recall, f1-score and confusion matrix. We get the highest accuracy with the decision tree classifier, which is 93.72%.
Analysis of Recognition Performance of Plant Leaf Diseases Based on Machine Vision Techniques
(Daffodil International University, 2022-03-01) Haque, Imdadul; Alim, Mohsin; Alam, Mahbub; Nawshin, Samia; Noori, Sheak Rashed Haider; Habib, Md. Tarek
Agriculture is the primary source of income for the majority of the population in Bangladesh. Agriculture is also a big part of the economy of the country. Therefore, it's more necessary to grow our crops and fruits and boost their harvests. Fruits are adored by the people of this country, and farmers love growing fruits. Owing to numerous diseases, both the quality and quantity of fruits are not meeting expectations. Native fruits are contracting many types of new diseases, and the magnitude of the problem is increasing alarmingly. To deal with this issue, quick detection of the disease and correct treatment or recuperation is required. In many cases, locals fail to even detect rare diseases. Thanks to the hug e advancement in technology, rare diseases can now be detected with the use of the right technologies. A good plant's growth is dependent on its leaves. Early leaf disease detection can help in keeping the leaves disease-free, as well as the plants and fruits. Our research focuses on identifying litchi leaf diseases by employing sophisticated image processing technologies to ensure the freshness of the leaves. A machine-vision-based technique, i.e., the Convolutional Neural Network (CNN), has been used in this research work.
Bangla song genre recognition using artificial neural network
(Scopus, 2024-06-24) Hasan, Md. Zahid; Akter, Mariam; Sultana, Nishat; Noori, Sheak Rashed Haider
p>Music has a control over human moods and it can make someone calm or excited. It allows us to feel all emotions we experience. Nowadays, people are often attached with their phones and computers listening to music on Spotify, Soundcloud or any other internet platform. Music Information retrieval plays an important role for music recommendation according to lyrics, pitch, pattern of choices, and genre. In this study, we have tried to recognize the music genre for a better music recommendation system. We have collected an amount of 1820 Bangla songs from six different genres including Adhunik, Rock, Hip hop, Nazrul, Rabindra and Folk music. We have started with some traditional machine learning algorithms having K-Nearest Neighbor, Logistic Regression, Random Forest, Support Vector Machine and Decision Tree but ended up with a deep learning algorithm named Artificial Neural Network with an accuracy of 78% for recognizing music genres from six different genres. All mentioned algorithms are experimented with transformed mel-spectrograms and Mean Chroma Frequency Values of that raw amplitude data. But we found that music Tempo having Beats per Minute value with two previous features present better accuracy.
Bayanno-net
(Proceedings of 2019 IEEE Region 10 Symposium, IEEE, 2019-06-07) Islam, Mohammad Shakirul; Foysal, Md. Ferdouse Ahmed; Noori, Sheak Rashed Haider
Handwritten digit recognition is one of the most novel topics from last few years. The complexity of recognition handwriting are differ in languages because of their shapes, character numbers and streak. Albeit Bangla is the 7th most popular language in order to the number of first language speakers. Remaining approaches use discrete feature expulsion methods and algorithms to recognize handwritten digits. Recently, Deep learning and convolutional neural network is used to solve the classification problem, it gives better accuracy for image classification with its distinct features. In this paper, we have proposed a Convolutional Neural Network referred as “ByannoNet”, to identify Bangla hand-written digits. We worked with the richest and popular dataset called NumtaDB generated and published by the Bengali.ai community. Our proposed model has achieved 97 percent accuracy with a very low cross-entropy rate.
Bengali Named Entity Recognition
(2019 10th International Conference on Computing, Communication and Networking Technologies (ICCCNT), IEEE, 2019-07-08) Rifat, Md Jamiur Rahman; Abujar, Sheikh; Noori, Sheak Rashed Haider; Hossain, Syed Akhter
Sequence labeling is a complex task in natural language processing where the data set used to be biased to a specific class mainly to “not named entity” class. Previously several machine learning approaches were harnessed for Bengali named entity recognition where additional information like Parts Of Speech (POS) tag, suffix value, optimal number of context words were required. This study aspires to give an overview of past methods on Bengali named entity task along with leveraging different neural networks on a new dataset at an easy way. A dataset consisting of 96697 tokens were annotated in the house where 67554 tokens were applied for training and 29143 words were for testing purposes. Then several deep learning methods were exploited where Bidirectional Gated Recurrent Unit (BGRU) came up victorious with f1 score 72.66%. The value may not be promising after comparing with other methods but different studies calculated the precision and recall value differently. Increasing the number of training data could raise the performance metrics along with other forms of word embedding.
Bengali Review Analysis for Predicting Popular Cosmetic Brand Using Machine Learning Classifiers
(Springer Nature Limited, 2022-11-14) Rabeya, Tapasy; Khatun, Eshita; Noori, Sheak Rashed Haider; Akter, Sharmin; Jahan, Israt
Nowadays, online platform has become one of the most popular media to express people’s thought of all ages. That made the online platform a precious source for getting almost every kinds of information. As online shopping is rising in no time in recent years, as a result millions of comments are generating every single day. These users generated opinions on social media and different websites has made it easier for the people choosing the right product for them. Hence, sentimental analysis is a sought-after research topic nowadays. Our research paper has portrayed an experimental study on different cosmetics products review. To do so, we have selected ten popular cosmetic brands for analyzing their product review and chosen to analyze Bengali comments or sentences. The main focus of our work was to get out the most popular cosmetic brands among ten chosen brands. We have applied four classification algorithm such as naive Bayes, random forest, decision tree, and support vector machine for analyzing the final outcome and found vaseline and clear are the most popular brands.
CSV-ANNOTATE: Generate annotated tables from CSV file
(IEEE, 2018-06-28) Mahmud, S M Hasan; Hossin, Md Altab; Jahan, Hosney; Noori, Sheak Rashed Haider; Bhuiyan, Touhid
The Semantic Web is a part of the current World Wide Web (WWW), which can facilitate a common mechanism to publish, share, and reuse data beyond the boundaries of web applications. It is widely believed that the majority of the datasets stored on the current web are in tabular data format (CSV, spreadsheets, SQL dumps, HTML tables etc), commonly in the comma-separated values (CSV) format. In order to prepare the CSV data semantically structured, interoperable, accessible and reusable for various web applications, they need to be extracted from the CSV files and converted into annotated table. Therefore, we propose an effective approach to generate annotated tables from CSV file. However, annotated table for CSV provides possibilities for data publishers to refer data validating, converting, displaying and inputting by following the Semantic Web standard. This research presents the conversion strategies of CSV file into annotated tables. Here, we design a parsing algorithm and development techniques to demonstrate the annotated tabular data model (column, row, and cell). An experiment is carried out to observe and compare the time efficiency of the annotation process. This method and findings provide a valuable reference for potential implementers to further operate the Semantic data.
Csv2rdf: Generating rdf data from csv file using semantic web technologies
(Journal of Theoretical and Applied Information Technology, 2018-10-31) Mahmud, S M Hasan; Hossin, M.A.; Jahan, H.; Noori, Sheak Rashed Haider
Recently, a large amount of Governments and public administrations data are stored on the Web in various file formats, mostly in the tabular data form such as Comma Separated Values (CSV) or Excel. CSV format is simple and practical, but it is difficult to express the relevant metadata such as data provenance, meaning of data fields, relationships between data fields, and user access approaches/rights, etc. In order to make the CSV data semantically structured, interoperable, accessible and reusable for various Web applications, they need to be extracted from the CSV files and converted into the Resource Description Framework (RDF) format that provides superior data assimilation and query functionality. In this paper, we focus on how the Semantic Web technologies are used to convert CSV data into RDF. Therefore, we present a method and techniques to parse the CSV file; the parsed CSV data are complemented with metadata annotations to generate the annotated tabular data model which is then converted into RDF triples. According to the conceptual correspondences between the CSV data model and RDF data model, we designed a set of algorithms to generate RDF triples from the CSV data. Our developed prototype tool, CSV2RDF, is used for evaluating the performance of the proposed method through real-world CSV datasets. The implementation and experimental outcomes demonstrate that our pro-posed method is feasible to generate RDF data from CSV datasets, with satisfactory performance on any size of data sets.
Efficacy and Acceptance of Virtual Classrooms During COVID-19
(Daffodil International University, 2021-06-13) Halder, Nabarun; Islam, S. M. Rakibul; Hosain, Md. Sarwar; Ahmed, Eshtiak; Islam, Ashraful; Noori, Sheak Rashed Haider
The sudden spread of COVID-19 shut down educational institutions worldwide, and Bangladesh was no exception. Educational institutions were forced to start their activities online; there was no alternative to keep the students in the study. Although online education has been seen as part of a futuristic approach, its effectiveness and acceptability still remains questionable when it comes to institutional education. It's yet to be investigated if online education can be as effective as contact teaching. We conducted an online survey to determine what students feel about online classes, how they accepted online classes, and how useful it was for them depending on their current situation. Our survey was open to everyone who has taken online classes during COVID-19, and the number of participants in our survey was 210. The survey provided with both qualitative and quantitative data which were then categorized into themes for analysis. Findings suggest, that most students feel that by rethinking the class style, if teachers can provide well-structured lecture content and have an equal focus on all students, it can be an alternative for them during emergency days.
Efficacy and Acceptance of Virtual Classrooms during Covid-19
(2021 3rd International Congress on Human-Computer Interaction, Optimization and Robotic Applications (HORA), IEEE, 2021-08-25) Halder, Nabarun; Islam, S. M. Rakibul; Hosain, Md. Sarwar; Ahmed, Eshtiak; Islam, Ashraful; Noori, Sheak Rashed Haider
The sudden spread of COVID-19 shut down educational institutions worldwide, and Bangladesh was no exception. Educational institutions were forced to start their activities online; there was no alternative to keep the students in the study. Although online education has been seen as part of a futuristic approach, its effectiveness and acceptability still remains questionable when it comes to institutional education. It's yet to be investigated if online education can be as effective as contact teaching. We conducted an online survey to determine what students feel about online classes, how they accepted online classes, and how useful it was for them depending on their current situation. Our survey was open to everyone who has taken online classes during COVID-19, and the number of participants in our survey was 210. The survey provided with both qualitative and quantitative data which were then categorized into themes for analysis. Findings suggest, that most students feel that by rethinking the class style, if teachers can provide well-structured lecture content and have an equal focus on all students, it can be an alternative for them during emergency days.
Empirical Study of Computational Intelligence Approaches for the Early Detection of Autism Spectrum Disorder
(Scopus, 2021) Khatun, Mst. Arifa; Ali, Md. Asraf; Ahmed, Md. Razu; Noori, Sheak Rashed Haider; Sahayadhas, Arun
The objective of the research is to develop a predictive model that can significantly enhance the detection and monitoring performance of Autism Spectrum Disorder (ASD) using four supervised learning techniques. In this study, we applied four supervised-based classification techniques to the clinical ASD data obtained from 704 patients. Then, we compared the four machine learning (ML) algorithms performance across tenfold cross-validation, ROC curve, classification accuracy, F1 measure, precision, recall, and specificity. The analysis findings indicate that Support Vector Machine (SVM) achieved the uppermost performance than the other classifiers in terms of accuracy (85%), f1 measure (87%), precision (87%), and recall (88%). Our work presents a significant predictive model for ASD that can effectively help the ASD patients and medical practitioners.
Empirical Study of Computational Intelligence Approaches for the Early Detection of Autism Spectrum Disorder
(Springer, 2020-09-30) Khatun, Mst. Arifa; Ali, Md. Asraf; Ahmed, Md. Razu; Noori, Sheak Rashed Haider; Sahayadhas, Arun
The objective of the research is to develop a predictive model that can significantly enhance the detection and monitoring performance of Autism Spectrum Disorder (ASD) using four supervised learning techniques. In this study, we applied four supervised-based classification techniques to the clinical ASD data obtained from 704 patients. Then, we compared the four machine learning (ML) algorithms performance across tenfold cross-validation, ROC curve, classification accuracy, F1 measure, precision, recall, and specificity. The analysis findings indicate that Support Vector Machine (SVM) achieved the uppermost performance than the other classifiers in terms of accuracy (85%), f1 measure (87%), precision (87%), and recall (88%). Our work presents a significant predictive model for ASD that can effectively help the ASD patients and medical practitioners.
Enhancing Sentiment Analysis using Machine Learning Predictive Models to Analyze Social Media Reviews on Junk Food
(Daffodil International University, 2023-12-20) Ajmain, Moshfiqur Rahman; Khatun, Mst. Farhana; Bandan, Sheikh Sadi; Rejuan, Arifur Rahman; Ria, Nushrat Jahan; Noori, Sheak Rashed Haider
In the last few years, the Use of social media has increased immensely. People share different types of opinions on social media like Facebook posts, comments, tweets etc. Sentiment analysis involves the process of categorizing these opinions. The aim of this study, find out the customer’s attitudes toward the restaurant. Nowadays sentiment review is gaining grip. The benefits of this sentiment analysis for restaurants is how customers like their food and as a result, the business of Bangladeshi restaurants will be more developed. The study focuses primarily on customers’ behavior, tastes, preferences, conversations, reviews, and objections. For this purpose 500 data are collected. There are six attributes in the dataset and based on customer reviews they are satisfied or unsatisfied. This exploration uses different classifiers of ML to develop review analysis like SVM, Random Forest, K-nearest neighbors, Decision Tree, Logistic Regression and XGBoost Classifier. And Comparing these algorithms’ performances, XGBOOST gives the greatest accuracy which is 83%.
FinTech: Deep Learning-Based Sentiment Classification of User Reviews from Various Bangladeshi Mobile Financial Services
(Springer Nature Limited, 2023-07-24) Ryan, Abdullah Al; Mahmud, Md. Shihab; Mahi, Hasibul Hasan Chowdhury; Hossen, Md Shakil; Shimul, Nazmul Islam; Noori, Sheak Rashed Haider
Banking has become an integral part of our lives. Fintech (Financial Technology) skyrocketed the number of people willing to use Mobile Financial Services (MFS) for their daily financial transactions. The banks are providing their services via mobile applications, which can be found on the Google Play Store. These Mobile Financial Services (MFS) provide mobility and increase efficiency by 10-fold. With an astonishing number of users came an abundant number of reviews for these apps. User reviews are the backbone of an application’s success. They provide information about hands-on experience. This study mainly focuses on the reactions of the users of such apps. Sentiment analysis is being used to draw out emotions from the users based on their written reviews. The primary goal of this paper is to examine the points of view of such application users. A total of 5414 pieces of data were collected from the Google Play Store and classified as negative, neutral, or positive. The data model has been evaluated using CNN, LSTM, and BiLSTM algorithms. Compared to CNN and LSTM, the BiLSTM algorithm produced the best model with an accuracy of 97.07%.
Machine Learning Technique Based Fake News Detection
(IEEE, 2023-01-15) Sutradhar, Biplob Kumar; Zonaid, Md.; Ria, Nushrat Jahan; Noori, Sheak Rashed Haider
False news has received attention from both the general public and the scholarly world. Such false information has the ability to affect public perception, giving nefarious groups the chance to influence the results of public events like elections. Anyone can share fake news or facts about anyone or anything for their personal gain or to cause someone trouble. Also, information varies depending on the part of the world it is shared on. Thus, in this paper, we have trained a model to classify fake and true news by utilizing the 1876 news data from our collected dataset. We have preprocessed the data to get clean and filtered texts by following the Natural Language Processing approaches. Our research conducts 3 popular Machine Learning (Stochastic gradient descent, Naïve Bayes, Logistic Regression,) and 2 Deep Learning (Long- Short Term Memory, ASGD Weight-Dropped LSTM, or AWD-LSTM) algorithms. After we have found our best Naive Bayes classifier with 56% accuracy and an F1-macro score of an average of 32%.
NN at CheckThat! 2023: Subjectivity in News Articles Classification with Transformer Based Models
(CEUR Workshop Proceedings, 2023-08-31) Dey, Krishno; Tarannum, Prerona; Hasan, Md. Arid; Noori, Sheak Rashed Haider
The CheckThat! Lab is a challenging lab designed to address the issue of disinformation. We participated in CheckThat! Lab Task 2, which is focused on classification of subjectivity in news articles. This shared task included datasets in six different languages, as well as a multilingual dataset created by combining all six languages. We followed standard preprocessing steps for Arabic, Dutch, English, German, Italian, Turkish, and multilingual text data. We employed a transformer-based pretrained model, specifically XLM-RoBERTa large, for our official submission to the CLEF Task 2. Our results were impressive, as we achieved the 1st, 1st, 2nd, 5th, 2nd, 2nd, and 3rd positions on the leaderboard for the multilingual, Arabic, Dutch, English, German, Italian, and Turkish text data, respectively. Furthermore, we also applied BERT and BERT multilingual (BERT-m) models to assess the subjectivity of the text data. Our study revealed that XLM-RoBERTa large outperformed BERT and BERT-m in all performance measures for this particular dataset provided in the shared task.