Aggression detection refers to the identification of offensive or harmful expressions in online communications and is crucial due to its significant societal implications and impact on online safety. However, detecting aggression in Bengali texts is particularly challenging due to the language’s complex morphology and the scarcity of annotated datasets and computational tools. Although the task has been extensively studied in high-resource languages, research in the resource-constrained Bengali language remains limited. To address the gap, in this paper, we developed a new dataset BOLT (Bangla Offensive and Lethal Texts) consisting of 4027 annotated instances classified into four classes based on severity of aggression- no aggression, hate speech, vandalism, and atrocity. To optimize model performance, we proposed a modified attention mechanism within transformers. Unlike conventional self-attention, which computes pairwise interactions between all tokens and treats their importance uniformly, our approach introduces a learnable token-wise weight matrix to assess the relevance of each token for aggressive text classification. The attention scores are obtained by multiplying the transformer’s token embeddings with this matrix to generate task-specific weights that better capture contextual significance. Moreover, we utilized a weighted ensemble of the optimized transformer models to combine their individual strengths. Experimental results show that our proposed approach outperforms existing methods and baseline models, achieving both accuracy and weighted F1-score of 87

Internet communication is now an integrated part of daily life with lots of communications related applications in computer systems. The proposed mobile communication application, named Converso, is an Android chatting application for real-time text chat with additional features like the live streaming and an artificial intelligence-based chatbot. Java is used in developing the back end of the application while XML has been used for the design aspect of the application. The application makes use of Firebase as the data storage as well as data synchronization solution. In this paper, the author describing the process of devising of Converso and its purpose of offering cost effective rapid communication channel for business promotions and education get Total: The paper focuses on describing the creation of Converso and indicating the main purpose of the creation of the service. Furthermore, it emphasizes the measures taken to safeguard user security within the context of existing trends of mobile communicating technologies.

Eye diseases, a significant global health concern, require timely detection to prevent vision loss. The alarming prevalence of eye diseases necessitates immediate action through early diagnosis, making it urgent to develop an automatic detection system. Many researchers have been working to develop such systems. Yet, existing solutions still face difficulties in achieving high-performance accuracy due to challenges like lacking feature effectiveness, high computational demands, and incomplete disease coverage. To overcome these challenges, we proposed a novel eye-disease detection system leveraging multi-stage deep learning technologies. In the study, we employed a preprocessing approach to ensure the system’s robustness against rotation and translation, enhancing its effectiveness across varied conditions. Then, we employed a lightweight three-stage deep learning approach for extracting effective features and specific advantages. In the procedure, Stage 1 focuses on extracting fine-grained features using deep learning layers where the layers can automatically learn and identify complex patterns associated with various eye diseases, improving feature effectiveness and overall system accuracy. Then, we employed stage 2, which is constructed with two branches, each composed of convolutional blocks and identity blocks; this stage extracts hierarchical features by concatenating the outputs of the two branches. This hierarchical approach captures both low-level and high-level features, enhancing the extracted features’ richness and robustness and leading to better classification performance. We concatenated the two branch features that fed into the classification module, producing a probabilistic eye disease presence map. By converting hierarchical features into precise disease predictions, this stage ensures accurate probabilistic outputs, aiding better decision-making and diagnosis. We evaluated the proposed model with OCT2017, Dataset-101, and Retinal OCT C8 datasets, demonstrating an accuracy improvement of up to 1% over existing state-of-the-art models in both multi-class and binary classification tasks. The lightweight design and reduced computational requirements of the model highlight its applicability for real-world deployment, particularly in resource-constrained environments. This computer-aided detection system offers a meaningful advancement in the field of automatic eye disease detection by providing a more accurate and efficient tool that can be deployed widely.

Traffic sign recognition is a critical aspect of intelligent transportation systems that enhances road safety, traffic management, and driver assistance. Notably, while significant research efforts have been devoted to many prevalent traffic signs, Bangladeshi sign recognition remains relatively underexplored due to a lack of traffic sign knowledge and the unawareness of the sector individuals. The scarcity of Bangladeshi traffic sign datasets has exacerbated this limitation, hindering progress. Moreover, most Bangladeshi traffic sign research predominantly relies on a limited number of signs, leading to diminished accuracy and the inability to detect all traffic signs. Additionally, existing systems grapple with suboptimal performance accuracy and heightened computational complexity, further emphasizing the research gap. To address these formidable challenges, we propose a robust Bangladeshi traffic sign recognition system using a deep learning-based CNN architecture. The main aim is to create a lightweight model that helps identify traffic signs correctly with less computational effort. We used the effective combination of various deep learning layers in the proposed system. A significant contribution of our work lies in creating a novel Bangladesh Traffic Sign Recognition (BDTSR) dataset, addressing the scarcity of data in this domain. This dataset comprises comprehensive information, including 48 types of traffic signs, and these images were taken from different roads in the country. Our dataset aims to fill a critical gap in Bangladeshi traffic sign research and provides a solid foundation for more extensive and inclusive studies in the field. Through this innovative approach, we aim to contribute significantly to the field of BDTSR, filling the gaps in traffic sign recognition and bolstering the accessibility of traffic signs within the Bangladeshi individual traffic community and beyond. Our evaluation of the newly created BDTSR dataset and the existing two benchmark datasets resulted in high recognition accuracies of 99.34\%, 96.97\% and 98.82\% accuracies for the BDTSR, German Traffic Sign Recognition Benchmark (GTSRB) and the Belgian Traffic Sign Dataset (BTSC) respectively. These resultshighlight the superiority of our model in the Bangladeshi Traffic Sign Recognition domain, outperforming existing models in terms of accuracy and computational efficiency.

As they play a significant role in autonomous driving and traffic safety, traffic sign identification and recognition have recently emerged as one of the most significant fields in image processing and computer vision. Early studies in this field offered several deep learning-based methods for classifying distinct traffic signs using various standard datasets. However, not many researchers focused on creating a dataset of traffic signs in Bangladesh and applying deep learning techniques to recognize them. In this research, we compare and contrast several deep learning models for recognizing traffic signs from the perspective of Bangladesh. We construct a novel dataset with over 2000 images representing thirteen distinct kinds of typical traffic signs in Bangladesh. Using data augmentation, about 8386 images are generated from the original dataset. Subsequently, transfer learning and fine-tuning approaches are applied to nine different deep learning models using this dataset, and the outcomes are compared. Results indicate that ViT had the highest validation accuracy of 99.91

Among other motif finding algorithms Planted Motif Search (PMS) or (ℓ, d) motif search algorithms are considered as next-generation motif finding algorithms. Here, n strings and two integer ℓ and d are provided as input in PMS. It gives outputs of all sequences of length M in each input string, where each occurrence varies from Min at most d points. This paper proposes a new algorithm, for the (ℓ, d) motif discovery problem in which we find all strings of length ℓ that seems in every string of a provided set of strings with at most d mismatches has been. After proper investigation of the PMS problem, it has been shown that this is an NP-hard exact algorithm. Usually, in the worst-case scenarios, all the known exact algorithms for PMS need exponential time in some of the underlying parameters. In this paper, we proposed a faster approach that reduces the searching time in the sample-driven part of the algorithm. In particular, we used dynamic programming techniques to eliminate the recalculation of the values of some common subtree and introduced some new speedup techniques like using a linked list, reduced ℓ-mers for making the algorithm faster. Due to the use of Linked List as a main speedup technique, we named our proposed algorithm as Linked List Implementation technique of PMS8 (LL-PMS8).

Clustering is one of the widely used techniques to find out patterns from a dataset that can be applied in different applications or analyses. K-means, the most popular and simple clustering algorithm, might get trapped into local minima if not properly initialized and the initialization of this algorithm is done randomly. In this paper, we propose a novel approach to improve initial cluster selection for K-means algorithm. This algorithm is based on the fact that the initial centroids must be well separated from each other since the final clusters are separated groups in feature space. The Convex Hull algorithm facilitates the computing of the first two centroids and the remaining ones are selected according to the distance from previously selected centers. To ensure the selection of one center per cluster, we use the nearest neighbor technique. To check the robustness of our proposed algorithm, we consider several real-world datasets. We obtained only 7.33%, 7.90%, and 0% clustering error in Iris, Letter, and Ruspini data respectively which proves better performance than other existing systems. The results indicate that our proposed method outperforms the conventional K means approach by accelerating the computation when the number of clusters is greater than 2.

Chronic kidney disease (CKD) is heterogeneous disorders that affects the renal functions and structures of millions of people around the globe, and it is one of the leading causes of morbidity and mortality. Given the circumstances, several studies had been conducted in order to detect CKD at an early stage. However, each of these studies has its own set of limitations such as the failure to employ proper methods for coping with missing values, anomalies, and class imbalance problems, overfitting issues, and so on. Taking into account the shortcomings that recent research has uncovered, we propose a novel CKD diagnosis method based on 1D Convolutional Neural Network (1D CNN) that overcomes the aforementioned drawbacks while also significantly improving diagnosis accuracy. The Chronic Kidney Diseases Dataset from the UCI Machine Learning Repository has been used in this study. MissForest imputation, a precise non-parametric missing value imputation process, has been used to handle missing data. Additionally, memory-efficient Isolation Forest has been applied to deal with anomalies. After evaluating the model with chronic kidney disease dataset, our proposed model achieved 99.21 % accuracy which is better than the state of the art method.

Planted Motif Searching (PMS) is considered as one of the most vital tool for finding all pair of motif on a genome sequence. In this problem, n length of string along with two integer ℓ and d are given as input and need to generate all possible sequence of length m, where each occurrence varies from at most d points. Generating Common d-Neighborhood is a vital step to find the planted motifs in every version of PMS algorithm. Particularly, in this step, we find the set of d-neighborhood between two ℓ-mers if the sufficient pruning conditions are fulfilled. In this paper, we have analyzed the previous algorithms for finding d-neighborhood and proposed a faster approach that reduces the searching time in the sample driven part of the algorithms. In particular, we have used dynamic programming techniques to reduce the repetitive calculation of the values of common subtrees and used advanced data structures to make the algorithm more time efficient.

In the case of finding the topological ordering of a directed acyclic graph (DAG), kahn’s and Depth First Search (DFS) topological sorting algorithms are used. Both of these algorithms time complexity is O(|V| + |E|). Here a topological sorting algorithm is proposed that is completely new and it reduces the time complexity of the previous algorithms. By separating the vertices having outgoing edges and the vertices having no outgoing edges then removing outgoing edges step by step, we can find a topological ordering of any DAG. The time complexity after using the proposed algorithm reduces to O( ∑_i = 1^| v || ( NE)_i|) for both average case and worst case but for best case it reduce to O(|V|), here |V| is the number of vertex and |(NE)| is the number of vertex contains at least one outgoing edge. This algorithm also can detect cycle in a graph. This algorithm cab be used for resolving dependencies, scheduling system, planning and in many graph algorithms.

Recently, many multifactor remote user authentication schemes have emerged to cover-up the security weaknesses of single factor user authentication systems. Among these schemes, the two-factor multi-server user authentication scheme has drawn a considerable amount of attentions of researchers. Subsequently, many two-factor user authentication schemes for multi-server architecture have been introduced in recent past. However, they were unable to prevent the security vulnerabilities like password guessing attack, user impersonation attack, privileged insider attack, server masquerading attack, denial of service attack, replay attack, etc. To address these security deficiencies, we propose a robust two-factor remote user authentication scheme for multi-server architecture using smartphone and biometric that can prevent all the major security vulnerabilities shown by existing schemes. We use ProVerif to demonstrate that our scheme fulfils all the required security properties. Through a comprehensive heuristic security and performance analysis, we show that our proposed scheme can overcome drawbacks of existing systems.

Mobile Crowdsourcing System (MCS) has emerged as an effective method for data collection and processing. In this paper, a brief discussion of the concept of Mobile Crowdsourcing system has been given where the main criteria of MCS is followed. The government or the census bureau performs the role of end user, the internet provider and some monitoring supervisor performs the role of service provider and the smart phone users can perform the role of worker. The whole country has been divided into some regions for counting the population, each region has been divided into several sub-regions. There will be a supervisor in each sub-region with a number of selected workers who will be checked on their own reliability and authentication on the basis of verification of personal information. An authenticated worker can collect information from a sub-region and the supervisor is able to determine the location of the worker. After collecting information, redundancy is checked using National Identity or birth registration number and stored after completing the verification process and used to make statistics including total population, population density, rate of birth, rate of death, rate of literacy etc. Population and household census process is a more important issue for the country. So, a census system or model has been proposed and designed for performing the whole census process with more authentication that reduce cost and time and make a faster calculation.

Quick Sort is considered as the fastest sorting algorithm among all the sorting algorithms. The idea of selecting a pivot was introduced in classical Quick Sort in 1962. This sorting algorithm takes favorably less time compared to other methods. It needs a complex time O( $$ nlogn $$ ) for the best case and O( $$ n^{2} $$ ) for worst-case which occurs when the input array is already sorted or reversely sorted. To reduce the worst-case complexity we provide a strong algorithm where it makes fewer comparisons and the time complexity after using this algorithm becomes a function of logarithm O( $$ nlogn $$ ) for worst-case complexity. We experimentally evaluate our algorithms and compare them with classical algorithms and with other papers. The algorithm presented here has profound implications for future studies of handling worst-case complexity and may one day help to solve this occurrence of the fastest sorting method.

Solving quadratic equation efficiently is a real-world challenge nowadays, due to its wide applications in the task of determining a product’s profit, calculating areas or formulating the speed of an object. The general approach of finding the roots of a quadratic equation is not enough efficient due to the requirement of high computation time. Because of the Genetic Algorithm’s stochastic characteristics and efficiency in solving problems it can be used to find roots of quadratic equation precisely. In modern athletics reducing the computation time of solving the quadratic equation has been so inevitable where using a genetic algorithm can find a quick solution that doesn’t violate any of the constraints and with high precision also. Optimization has been done in the Crossover and Mutation process which has reduced the number of iterations for solving the equation. It reduces the time complexity of the existing approach of solving the quadratic equation and reaches towards the goal efficiently.

In developed countries, Alzheimer's disease (AD) is one of the major causes of death. Until now, clinically there is not have any diagnostic method available but from a research point of view, this disease detection accuracy is produced by computational algorithms. There are many researchers who are working to find about Alzheimer's disease property, its stages, and classification ways. This research plays a vital role in clinical tests for medical researchers and in the overall medical sector. One of the major problems found by the researchers in the field is the large data dimension. In the study, we proposed an efficient dimensionality reduction method to improve Alzheimer's dis-ease (AD) detection accuracy. To implement the method first we cleaned the dataset to remove the null value and removing other unacceptable data by some preprocessing tasks. On the preprocessed data first, we have split into training and test dataset then we employed a dimension reduction method and there-fore applied a machine learning algorithm on the reduced dataset to produce accuracy for detecting Alzheimer's disease. To overserve and calculate the accuracy we computed confusion matrix, precision, recall, f1-score value and finally accuracy of the method as well. Reducing the dimension of data here we applied consequently Principle component analysis (PCA), Random Projection (RP) and Feature Agglomeration (FA). On the reduced features, we have applied the Random Forest (RF) and Convolution neural network (CNN) machine-learning algorithm based on the dimensionality reduction method. To evaluate our proposed methodology here we have used Alzheimer's disease neuroimaging initiative (ADNI) dataset. We have experimented for (i) Random forest with principal component analysis (RFPCA), (ii) Convolution neural network (CNN) with (PCA) (CNNPCA), (iii) Random forest with Ran-dom projection(RFRR), and (iv) Random forest with Feature agglomeration(RFFA) to differentiate patients with the AD from healthy patients. Our model namely Random forest with Random projection(RFRP) has produced 93% accuracy. We believe that our work will be recognized as a groundbreaking discovery in this domain.

The classical algorithm which is dedicated to resolve job sequencing problem with a deadline (JSD) needs exponential time O( n^2 ), where sorting algorithm [O( nlog( n ) )-(Merge Sort)] must have to use to sort all the jobs in decreasing order of their profit and it is a greedy technique. To reduce the complexity of this classical algorithm, we nullify the sorting algorithm using dynamic programming approach in the proposed algorithm. The time complexity after using this approach reduces to O( mn ), where no sorting algorithm [O( nlog( n ) )-(Merge Sort)] needed, which has been shown by proper explanation. Here, we were also given a novel approach to resolve the job sequencing problem using Dynamic Programming and it is a unique approach that always finds an optimal solution. By using this approach, a proper algorithm has been developed in this paper. Besides, finding maximum profit and the sequence of the job to obtain maximum profit, this algorithm gives the sequence of jobs for a specific profit or near a specific profit.

We proposed a classification technique of EEG motor imagery signals using Logistic regression and feature extraction algorithm using filter bank common spatial pattern (FBCSP). Main theme of FBCSP is that the signals decomposed into 5 sub band then calculated CSP for each sub band, this algorithm also allows automated frequency band selection. We combined each subband CSP feature vector, feed this feature vector into machine learning algorithm. In the paper Logistic regression is used to classify among multiple classes. To evaluate this method, we used here publicly available dataset namely Brain-Computer Interface competition IV-2a. Because of high accuracy and kappa that shown in accuracy table that proposed method is promising.

Agriculture is the foundation factor of many developing countries' economy. Now we are in the era of industrial revolution and especially the developing countries are shifted their vision completely to this revolution. But, they are not thinking about smart cultivation and more production from this ancient key factor of economic growth. Still farmers are following the ancient techniques for cultivation without thinking different dimensions or key factors for better production. Farmers should understand the present context for cultivating any crops to get proper profit against any specific cultivation. And to do this study they have to analysis statistical data of production report over a long period considering climate, region, availability of pesticides, period specific demand and so many dimensions. So we have to employ the context-aware computing to estimate the reasonable balanced production of any crops. In many countries like; Bangladesh in every cultivation season many farmers are not getting proper price of their crops due to over production. Sometimes farmers fail to return their invested money for their cultivation where their time and physical effort is completely a loss. To address this vital issue our research work will do proper context-aware analysis. This paper's agricultural context-aware computing for balanced production deals with extract-transform-loading (ETL), data warehousing (DWH), dimension (period, weather, pesticides etc.) specific data visualization, and supervised learning process to estimate future production.

DNA sequence analysis & comparison computation is a vital task in terms of memory & time which is used huge size of data set for biological research.Perfectly aligned sequence find out the matching point or mismatches between two sequences.Our proposed algorithm is composed of two major part.The first part is Fast Splitting(FS), a "Recursive technique" based algorithm which divides the source sequence in appropriate and exact length according to the preference of target sequence.Second part is Fast_Maximum Matches Subsequence Finder(Fast_MMSS).It builds the specialized successor table according to the identical characters of two strings (TLSS & Target Sequence).Then using some special pruning condition, we get the final MMSS.In previous work, dynamic programming and some sorts of Brute force techniques are applied which are faster in terms of time but requires huge memory, while our proposed algorithm maintains the "Time and Space" tradeoff.

Password security can no longer provide enough security in the area of remote user authentication. Considering this security drawback, researchers are trying to find solution with multifactor remote user authentication system. Recently, three factor remote user authentication using biometric and smart card has drawn a considerable attention of the researchers. However, most of the current proposed schemes have security flaws. They are vulnerable to attacks like user impersonation attack, server masquerading attack, password guessing attack, insider attack, denial of service attack, forgery attack, etc. Also, most of them are unable to provide mutual authentication, session key agreement and password, or smart card recovery system. Considering these drawbacks, we propose a secure three factor user authentication scheme using biometric and smart card. Through security analysis, we show that our proposed scheme can overcome drawbacks of existing systems and ensure high security in remote user authentication.