Browsing by Author "NUSRAT JAHAN"

Now showing 1 - 3 of 3
  • No Thumbnail Available
    Item
    Development of Regression-Based Mathematical Models for Early Detection of Lung and Breast Cancer Using Clinical Parameters
    (2025-07-30) NUSRAT JAHAN
    Cancer diagnosis at an early stage is vital in enhancing better outcomes of treatment and prevention of fatalities, particularly in low- and middle-income economies, such as that of Bangladesh, where modernized diagnostic methods are less available. The proposed research aims to construct regression models to detect lung and breast cancer in their early stages using clinically and efficiently accessible diagnostic parameters. A study of 1,000 patients with lung cancer and 570 patients with breast cancer was conducted. The most important predictors were identified through correlation matrix analysis to perform the feature selection. The Response Surface Methodology (RSM) establishes both single-order and polynomial regression models to capture both linear and nonlinear correlations between variables. ANOVA was used to test the statistical significance and the predictive power of the models. In the case of lung cancer, factors like history of smoking, genetic risk, and chronic lung disease were identified as 12 significant predictors. In the case of breast cancer, characteristics such as radius, perimeter, concavity, and compactness turned out to be influential. Separate datasets were used to verify the empirical validity and strength of the models. The findings indicate that the second-order polynomial models performed better than the linear models because they were able to capture the complex interaction of the variables. This study demonstrates that modeling based on regressions offers a cost-efficient, interpretable, and scalable approach to early cancer screening. These models present significant opportunities to support clinical decision-making in resource-constrained settings and facilitate early intervention in healthcare environments with limited resources. The innovative procedure has the potential to make a significant contribution to the improvement of cancer screening procedures and the current state of health in underserved communities, such as Bangladesh, by reducing dependence on costly diagnostic equipment. Keywords: Early detection, mathematical model, regression analysis, ANOVA, Response Surface Methodology (RSM), Lung cancer, Breast cancer.
  • No Thumbnail Available
    Item
    INTERFERENCE MANAGEMENT PROCEDURE IN DEVICE TO DEVICE COMMUNICATION TECHNOLOGY IN 5G CELLULAR NETWORK
    (2019-06-30) NUSRAT JAHAN
    Device-to-Device communication (D2D) process are used in 5G cellular network. D2D allows communication between two devices, without participation of the Base station or the evolved NodeB (eNB). In 5G cellular networks, with D2D communication enabled within is considered as two-tire networks. These two tire networks are referred to as macro-cell tire and device tire. The two-tier network needs to be designed with smart interference management strategies and appropriate resource allocation schemes. In two-tire network architecture cellular users and D2D users share common Resource Blocks (RBs). Such paradigms allow potential increase in the number of supported users but it increase the cost of interference. Because of this we need to design an efficient interference management system. Interference can be reduced by appropriate resource allocation and power control. We propose a two steps approach. In the first step we concerns with the efficient RB allocation to the users. The second one is the transmission power allocation. . We discuss about a minimum interference resource allocation algorithm based on the TDMA method and power control scheme. Firstly peer has to detect, then select the best mode (D2D or Cellular) depends on the distance. Two peer discover technique are used, one is fully dependent on the base station network and another is semi dependent on the network. We present an optimized mode selection process .For D2D users and cellular users resource are allocated, depends on the path gain. We proposed a minimum interference algorithm to allocate resource. In order to control power three power control scheme has been used. Hence in terms of increased number of users, interference reduction and power minimization can be done efficiently. After simulation using MATLAB we analysis the developed algorithm. We see that D2D communication is fast and efficient if we able to mitigate interference.
  • No Thumbnail Available
    Item
    UTILIZING TEXT SENTIMENT AUTOMATIC BENGALI BULLYING DETECTION THROUGH MACHINE LEARNING
    (2022-08-30) NUSRAT JAHAN
    The number of Bengali language users on social media is rapidly increasing, and the use of code-mixing and transliteration. Multiple languages mix using code-mixing. Transliterated Bengali allows writing Bengali with Latin words. Manually reviewing and removing bullying content from social media can be time-consuming, which is undesirable in today's technologically automated world. The machine learning approach can be convenient in keeping the system updated with new types of abuser approaches. Machine learning algorithms such as a k-nearest neighbor, Naive Bayes, Support Vector Machine, Random Forest, Logistic Regression, and AdaBoost classification were applied to distinguish transliterated Bengali cyberbullying content. The dataset contained not only transliterated Bengali text but also Bengali and Code-Mixed Bengali text. Baseline features from the social media dataset were used with sentiment and personality features. The features were the number of likes and dislikes, profane words, positive words, negative words, related to posts, and sentiment. TF-IDF and n-grams technique was used for feature extraction. The performances of the algorithms were evaluated utilizing precision, recall, accuracy, F1 score, AUC, and ROC curve. Among sentiment analysis. The result shows that the linear SVM algorithm achieved the highest accuracy of 64.224% to classify sentiment for test data. In contrast, the SVM and RF outperformed other classifiers in detecting bullying comments. For testing data, the SVM and RF achieved 94.83% and 94.40% accuracy, respectively. This work will have an impact on reducing transliterated and code-mixed Bengali cyberbullying.