Search

Sundarban, the world biggest mangrove forest, was announced as a World Heritage in 1997 and Ramsar Site in 2007. It's 62% lies in Bangladesh (Khulna, Bagerhat, Satkhira, Pirojpur, and Barguna Regions) and 38% in India, which become a vanguard for coastal communities. This ecosystem's floral and faunal structure incorporates 334 floral, 300 birds, 120 fish, 35 reptiles, and 32 mammals species. It contributes 41% of the total forest income of Bangladesh. This paper is intended to recognize the significant reasons for biodiversity calamity and its related effects on lives and degradation of the environment. A total of 105 published articles and reports were identified by searching four keywords: degradation, biodiversity, Sundarban, and Bangladesh. By considering the scope, 28 articles and 17 reports were incorporated to satisfy the objectives. The study revealed that the prominent causes of biodiversity loss of Sundarban might be due to intensified shrimp cultivation, increased dependency on forest and illicit felling, changing land use pattern, oil spillage, and pollution, forest fire, improper management practices, plant diseases, tourism activities, upstream withdrawal of river water, salinity intrusion, climate change, and some natural causes. The impacts were the loss of diversity of floral and faunal species subjected to regionally extinct, critically endangered, endangered, vulnerable, near threatened, and least concern condition. Occupational change and deterioration of the quality of life of surrounding communities are identified as impacts. Finally, the study recommended the formulation of a special mangrove forest policy and proper management practices of Sundarban with several environmental protection strategies that should be adopted to conserve this unique ecosystem.

Text classification has a growing interest among NLP researchers due to its tremendous availability on online platforms and emergence on various Web 2.0 applications. Recently, text classification in resource-constrained languages has been bringing much attention due to the sharp increase of digital resources. This paper presents a CNN based text classification model for one of the low resource languages like Bengali. The goal of the Bengali text classification is to assign a particular category to a text into one of the pre-defined categories based on its semantic and syntactic meaning. The proposed system comprises of four key modules: embedding model generation, Text to feature representation, training, and testing. The classification system trained and validated with 39, 079 and 6, 000 text datasets. Experimental evaluation with 9, 779 test datasets shows the accuracy of \(96.85\%\), which indicates the superior performance compared to the existing techniques.

Web text categorization is a procedure of deliberately assigning a web text document into one of the pre-defined classes or categories. It is a very challenging task to manipulate, organize, and categorize an enormous amount of web text data in manually. This paper proposes an automatic text categorization framework to classify Bengali web text data using deep learning. The proposed framework comprises of three key constituents: text to feature extraction, training, and testing. The categorization framework is trained, validated, and tested at 120K, 12K, and 36K datasets, respectively. The proposed system achieved \(99.00\%\) accuracy in the training phase, \(96.00\%\) in the validation phase, and \(95.83\%\) in the testing phase, respectively.

Eye cataract is a condition in which the lens of the eye becomes clouding or less transparent. This affects the clear vision and is the most prevailing causes of blindness. Therefore, early cataract detection and prevention may reduce the blindness rate and surgery pain of the patients. This paper presents an eye cataract detection system using Deep Convolution Neural Network (DCNNs) comprising two modules: training and testing. The proposed DCNNs architecture is trained, validated and tested with retinal fundus images. Experimental result shows that the proposed system is capable of detecting eye cataract with high accuracy.

In recent years, the amount of digital text contents or documents in the Bengali language has increased enormously on online platforms due to the effortless access of the Internet via electronic gadgets. As a result, an enormous amount of unstructured data is created that demands much time and effort to organize, search or manipulate. To manage such a massive number of documents effectively, an intelligent text document classification system is proposed in this paper. Intelligent classification of text document in a resource-constrained language (like Bengali) is challenging due to unavailability of linguistic resources, intelligent NLP tools, and larger text corpora. Moreover, Bengali texts are available in two morphological variants (i.e., Sadhu-bhasha and Cholito-bhasha) making the classification task more complicated. The proposed intelligent text classification model comprises GloVe embedding and Very Deep Convolution Neural Network (VDCNN) classifier. Due to the unavailability of standard corpus, this work develops a large Embedding Corpus (EC) containing 969 , 000 unlabelled texts and Bengali Text Classification Corpus (BDTC) containing 156 , 207 labelled documents arranged into 13 categories. Moreover, this work proposes the Embedding Parameters Identification (EPI) Algorithm, which selects the best embedding parameters for low-resource languages (including Bengali). Evaluation of 165 embedding models with intrinsic evaluators (semantic & syntactic similarity measures) shows that the GloVe model is more suitable (regarding Spearman & Pearson correlation) than other embeddings (Word2Vec, FastText, m-BERT) in Bengali text. Experimental results on the test dataset confirm that the proposed GloVe + VDCNN model outperformed (achieving the highest 96.96 % accuracy) the other classification models and existing methods to perform the Bengali text classification task.

Automatic document categorization has gained much attention by natural language processing (NLP) researches due to the enormous availability of text resources in digital form in recent years. It is the process of assigning a document into one or more categories that help the document manipulate and sort quickly. An efficient information processing system is required due to the rapid growth of Bengali text contents in digital form for searching, organizing, and retrieving tasks. In this paper, we proposed a framework for classifying Bengali text documents using deep convolution nets. The proposed framework consists of word embedding and document classifier models. Experiments with more than 1 million Bengali text documents reveals that the proposed system worthy of classifying documents with 94.96% accuracy.

The automated categorization of text documents into predetermined categories has witnessed a growing in the last few years, due to the huge availability of documents in digital form and the ensuing need to organize them. Automatic document categorization is the process of assigning one or more categories or classes to a document, making it easier to manipulate and sort. This paper proposes a Bengali document categorization technique based on word2vec word embedding model and stochastic gradient descent (SGD) statistical learning algorithm with multi-class svm. The semantic features of a document are extracting by Word2Vec and SGD improve the classification complexity with multi-class SVM that classify the unlabeled data. The experimental result with 10000 training and 4651 testing documents shows the 93.33% accuracy.

Word sense disambiguation is the process of identifying appropriate sense or meaning of a word in a sentence, when the word has multiple meanings. In this work, we propose a disambiguation system using Naive Bayes classifier and ANN to disambiguate the Bangla words. Naive Bayes classifier for training phase and using statistical information gained from the training dataset the unknown ambiguous words are predicted and gives detecting the correct sense using ANN. For the testing phase we used general regression neural network. The result of the evaluation shows that the proposed word sense disambiguation system can disambiguate the ambiguous Bangla word with 82% accuracy.