Browse
Recent Submissions
Item Floyd-warshall algorithm based on picture fuzzy information(UMT, Lahore, 2022) Aqsa MajeedThe revolution in technological aspects using information and communication technology and its involvement in almost every part of our socio-economic life span has been increased in last two decades with tremendous growth rate never seen before. Thus, the growing need of high end powerful computational model like Floyd-Warshall plays a vital role to accomplish computational node based needs in our daily life. Floyd-Warshall is the common algorithm used for finding the shortest path between any pair of nodes. These weights can be positive or negative. It works well for crisp weights, but the problem arises when weights are vague and uncertain. Let us take an example of computer networks, where network condition changes rapidly due to changes in a network situation, and the chosen path may not be suitable anymore. In computer networks or their related scenarios where only one parameter is not involved for deciding the optimal path between any pair of nodes. In this paper, we design a new variant of the Floyd-Warshall algorithm that finds an All-Pair Shortest Path (APSP) in an uncertain situation of a network. In the proposed methodology, multiple criteria and their mutual association may involve the selection of any suitable path between any two node points, and the values of these criteria may change due to an uncertain environment. We used trapezoidal picture fuzzy addition, score, and accuracy functions to find APSP. We also compute the time complexity of this algorithm and compare the proposed algorithm with the classical Floyd-Warshall algorithm and fuzzy Floyd-Warshall algorithm.Item Design of a cipher based on key dependent trigonometric substitution box(UMT, Lahore, 2022) YASIR BILALThis paper presents a cipher based on a chaotic substitution box which helps to secure the information. There are many components of cipher and one of the main components is the Substitution Box which is responsible for the chaos in the data. Information is a vital resource for any organization, which can be shared among multiple resources over some communication medium. While sharing through medium there is a chance of attack which may result in data loss. To secure this process, there is a need for an algorithm which is called a cipher. A cipher is responsible for securing the information with the help of a substitution box which produces chaos. In this research, the substitution box is formed with the help of trigonometric functions. The durability of the substitution box is analyzed by its properties. Furthermore, this paper introduces a new block cipher that encrypts 128 bits of data using a proposed substitution box. The cipher has 12 rounds of encryption process that takes 128 bits of key to perform multiple operations. The key will be generated for each round that will be used for the encryption. The cipher contains four major functions followed by the pre function. The proposed substitution box offers great complexity and the cipher provides a great encryption standard. The constructed substitution box has been analyzed with different metrics and the cipher has been tested as well.Item Diabetes prediction using machine learning technique(UMT, Lahore, 2022) SHAFQAT ULLAHDiabetes is a biological disorder that impacts people in almost every country and, if not discovered early, can lead to significant complications like stroke, kidney failure, and premature death. To combat this, a number of researchers are attempting to predict diabetes at an early stage using a variety of ways. Diabetes is diagnosed by a range of widely available traditional techniques based on physical and substance tests. Many people who have diabetes don‘t know about it. There are currently about 420 million people living with diabetes. According to World Bank data, the Marshall Islands has the highest diabetes prevalence rate, with 3:10 persons suffering from the disease. The extraction of diabetes feature information and the medical record for the diagnosis and treatment of sickness becomes more important to stimulate the growth of diabetes prediction and community medicine. Machine learning, data mining, artificial neural networks, fuzzy structures, genetic algorithms, rough collection, and various algorithms are among the methodologies and implementations for diabetes prediction [1]. Different state-of-the-art diabetes prediction approaches are given and compared in this research, as well as the machine learning techniques they used. Diabetes is a chronic disease so its prediction is necessary for saving lives. This survey paper helps to find out some previous literature limitation and help out to narrow down the importance of diabetes prediction. The major steps for prediction is data-preprocessing, feature selection, data splitting, apply various machine learning techniques to train data and get prediction results for disease data sets.Item Monitoring students’ engagement in class using facial emotion recognition(UMT, Lahore, 2022) MUHAMMAD ADNAN BASHIRSeveral recent. years have seen the introduction of Artificial Intelligence and Learning Analytics to the educational system. It has become one of the most important and difficult challenges for educators, researchers, and policymakers to devise a system that can determine the level of student engagement. This is because of the rise of distance learning in general and e-learning. Most of the previous work that has been done on the topic of facial expressions of emotion has concentrated on six primary categories, namely happy, surprised, angry, sad, afraid, and derisive. Consequently, the purpose of this paper is to perceive and analyze human emotions and affective states through Facial Emotion Recognition (FER) technology, and we also present a system to detect the engagement level of students. It uses only information provided by the typical built-in web-camera present on a laptop computer and was designed to work in real time. We combine information during classroom lectures about the movements of the eyes and head with facial emotions to produce a concentration index with two classes of engagement: active and inactive. This means that students are either paying attention (active or inactive) to the lecture or not. This paper can be used as a brief guide for those new to the field of FER, providing a general understanding the most recent state of the art experiments and fundamental knowledge, as well as for more seasoned researchers looking for promising directions for future work. However, humans regularly make much more use of emotional facial expressions. We compared the results of different methods to gain confidence in our model.Item An NLP based machine translation model for the deaf students of computer science of intermediate(UMT, Lahore, 2022) MARIAM MASHOOQThe deaf and hard-of-hearing people face many difficulties in almost every phase of their life whether they want to communicate with each other or want to learn education. The method they use to communicate with each other or with hearing people is known as sign language. Every country has its sign language according to its language and dialects, following the respective set of characters, vocabulary and general structures. This research work has proposed an intelligent system for the deaf and hard-of-hearing students that comprise a machine translation framework after implementing the core Natural Language Processing techniques. Unfortunately, as our country does not have proper medium, deaf and hard-of-hearing students always face a barrier in their education. By using these state-of-the-art NLP techniques, an avatar system has been created that translates a natural language into Pakistan Sign Language to facilitate the students so that they can learn the basic computer fundamentals and improve their knowledge in the domain of Computer science. The standardized curriculum of the Punjab Curriculum and Textbook Board has been used for the research work. The proposed machine translation model is able to work on word-level and sentence-level translation. This language model is a 3D avatar that translates the English language after understanding and analyzing all the features together with word orderings, words meanings, number features. HamNoSys notation has been used as it contains both manual and non-manual features. This notation provides the source language in the form of SiGML and then generates the streams of signs in the form of gestures like a human interpreter for the Pakistan Sign Language. This framework will perform as a robust medium for the hard-of-hearing students in Pakistan to learn applied subjects like computer science and provide them a way for choosing their path of interest in the technological field.Item Medical diagnosis based on single-valued neutrosophic information(UMT, Lahore, 2022) Maria AhmadWomen with heart disease during pregnancy are at higher risk, which can harm the fetus. This risk can be reduced if we diagnose and treat it early. The decision-making system is very helpful in such situations. Many clinical decision-making systems have been proposed, but they are too complicated for medical experts to understand and adapt. Here, we develop a new neutrosophic model for early diagnosis and explain it using explainable artificial intelligence. Our model is taking eight symptoms and signs as inputs and determines the diagnosis, type of treatment, and prognosis. Age, obesity, smoking, family pathological history, personal pathological history, electrocardiogram, ultrasound, and functional class are the inputs of this model. Six diagnoses can be made- obstruction at existing, obstruction at entry, rhythm disorder, conduction disorders, congenital diseases, genetic diseases. The types of treatments are pregnancy interruption, diuretic treatment, anti-arrhythmic treatment, treatment with beta-blockers and anticoagulants treatment. The prognosis is- eutectic delivery, dystocic delivery, the child with complications, child without complications, mother with complications, and mother without complications. The main parts of this system are neutrosophication, knowledge base, inference engine, de-neutrosophication, and explainability. To present the entire execution of the proposed system, we design an algorithm and compute its time complexity to demonstrate the working of the entire system. We compared the results of different methods to gain confidence in our model.Item A survey on blockchain acquainted software requirements engineering(UMT, Lahore, 2022) MISHAAL AHMEDRequirements are the basis of software development practices. Ambiguities in requirements leads a project to a point of failure or penalizes it with high budget and time for defect traceability. The ever-growing demand for advance computing systems has increased the complexity of Software Requirements Engineering (SRE) practices. Blockchain systems require specialized SRE practices as the issues of requirement traceability (RT), developer/client confidentiality, and requirement negotiation (RN) typically exist in conventional approaches, which require more improvement. Moreover, the blockchain technology incorporates the capacity to function as an infrastructure for SRE framework providing transparency, security, and reliability. Even though the significance of studying blockchain in context of SRE is evident, it is still in its infancy. To the best of our knowledge, none of the previous studies conducted a comprehensive research study based on survey in this domain. In this research, we have provided a comprehensive review on the aspects of blockchain acquainted SRE practices. We have presented SRE based quality improvement factors and outlined the need for blockchain technology in this domain. Furthermore, we have classified SRE practices based on blockchain engineering. In addition, we have proposed a generic SRE model built on blockchain infrastructure along with its workflows. Similarly, we have provided implementation guidelines for the future development guidance of SRE applications built on blockchain technology. Finally, we have presented the current research challenges and provided future directions based on blockchain acquainted SRE.Item Sarcasm detection in urdu using tweets data(UMT, Lahore, 2022) Shanzay GulAn essential task in sentiment analysis is the automatic detection of sarcasm in textual data. Because sarcastic remarks can convey the opposite meaning and depend on context, this situation is complicated. Due to the dearth of linguistic resources online, identifying sarcasm in comments written in Urdu text poses an even more challenging problem. In this research, we have performed sarcasm detection in Urdu using tweet data. We have used the iSarcasm Eval dataset which is a labeled dataset. We translated English tweets into Urdu tweets using Google Translator and then trained the model to perform sentimental analysis on it using Supervised learning to detect sarcasm. We used SVM and logistic regression using BOW and TF-Idf. We got the accuracy of the results 70%,67%, and 74% for logistic regression using BOW, logistic regression using TF-Idf, and SVM using BOW respectively.Item Protein carbonylation sites prediction by using deep neural networks(UMT, Lahore, 2022) SANIA SEHARPost-translational modifications (PTMs) are the processing events that can modify the properties of a protein in case of proteolysis or if any modifying group is added to one or more amino acids. Oxidative stress arises when there is an imbalance in the regulation and production of reactive oxygen species (ROS) and reactive nitrogen species (RNS). The protein carbonylation is a post-translational modification and a biomarker foroxidative stress because of some special traits it has such as its early development, irreversibility, and stability. Overabundance in external oxidative stress, obesity, and aging increase the density of protein carbonylation, and this enlargement signals early-stage diseases. There are many human diseases associated with protein carbonylation such as diabetes, chronic renal failure, sepsis, chronic lung disease, Parkinson’s disease, and Alzheimer’s disease. So, it is very important to identify protein carbonylation sites in pathology and physiology. It can provide important evidence for basic research and medication development. Carbonylation is vulnerable to only a subset of proteins and most of the carbonyl groups are made from four types of amino acid residues known as lysine (K), proline (P), arginine (R), and threonine (T). It is very costly and time-consuming to determine the carbonylation sites experimentally, especially in the case of broad datasets. So, computational methods are recommended for the recognition of carbonylation sites in proteins. This research develops a protein carbonylation sites predictor. An experimentally verified benchmark dataset is used for the study. Three deep neural networks are trained and tested; fully connected neural network, recurrent neural network, and convolutional neural network. Recurrent Neural Networks (RNNs) are trained with simple RNN units, Gated Recurrent Units (GRU), and Long Short-Term Memory (LSTM) units. The convolutional neural network provided the best performance with an accuracy value of 0.913. So, it can learn deep representations among data efficiently. This study has the novelty to use deep neural networks for carbonylation sites prediction. It is providing comparable results with state-of-the-art predictors of carbonylation sites. Only two models are providing better performance than the proposed model. I have not done any feature extraction and still, I got comparable results. All of the other previously proposed predictors have low performance than the proposed methodology. So, the proposed model is stable, valid, and reliable than all of these predictors. This study is providing a baseline to researchers who want to work with deep neural networks for carbonylation sites prediction.Item An intelligent model for the prediction of PIWI-interacting RNAs and their functions(UMT, Lahore, 2022) Anam UmeraA documented class of short non-coding RNA molecules is known as PIWI interacting RNA (Pi RNA). The creation of new drugs and the identification of various tumor types are linked to the Pi RNA molecules. Additionally, it is related to controlling transcription of genes, squelching transposons, and preserving genomic stability. The discovery of pi RNAs and their functionality has grown to be a significant research topic in bioinformatics as a result of the crucial influence that pi RNAs play in bio. The 2L-PiRNA-ML predictor, is a strong two-layer predictor that is suggested in this research to enhance the prediction of PiRNA and their functionality. The suggested model uses Quadratic Discriminant Analysis Classifier, Linear Discriminant Analysis, Passive Aggressive Classifier, Gradient Boosting Classifier, Extra Tree Classifier, Support Vector Machine, Logistic Regression, Random Forest, Ridge Classifier, Ridge Classifier CV, Bagging Classifier for classification. It also employs DNC and TNC for extraction of features. The suggested model is created using a two-layer construction strategy. The 1st layer makes a prediction about a given sequence whether it is Pi RNA or not, and the 2nd layer makes a prediction about a given PiRNA sequence whether it is having the function of instructing target mRNA dead enylation or not. Proposed model overall accuracy was 93.06 % at the first layer and 88.46 % at the second layer.Item A new cipher design based on trigonometric dynamic S-Box(UMT, Lahore, 2022) SYEDA UM-E-RUBAB TIRMAZIAn information security system is a vital need for the security of data. Strong cryptographic ciphers are required to ensure the comprehensive performance of information security systems. Ciphers are mathematical algorithms that convert plaintext into ciphertext and provide security to the data. Cipher uses different operations like XOR, shift, permutation operation, substitution operation, and many more. For any cipher, the substitution box plays a key role in strong security performance parameters. An S-Box is utilized in a security system that is based on the block ciphering technique, as it is the core element for providing non-linearity in a cryptosystem. In our study, we proposed an S-Box construction method for generating dynamic S-Box with cryptographically strong properties. The idea proposed in this study is inspired by chaotic maps. The proposed dynamic 8*8 S-Box is analyzed using the strict avalanche criterion (SAC), bit independence criterion (BIC), non-linearity, and differential uniformity. The proposed method for the construction of S-Box was proved to produce S-Boxes with high non-linearity and randomness characteristics and the proposed cipher is secure. The proposed cipher consists of ten rounds for encryption, decryption, and key generation. The master key used is 128 bits. The operations used in the proposed cipher are XOR operation, substitution operation, complement, row transformation, etc.Item Analysis and prediction of COVID 19 using big data analysis(UMT, Lahore, 2022) MIAN MUHAMMAD YASIRMany industries, from manufacturing and commerce to law enforcement and healthcare, can benefit from the IOT applications and smart sensors. These Internet of Things based appliances and sensors generate a wealth of information that, if studied using big data analytics, might prove extremely useful to healthcare providers. Human health, life, and productivity are under danger due to the current new coronavirus pandemic (COVID-19) epidemic. The pandemic was successfully countered with the use of Internet of Things and big data technology. Methods that may be used to achieve this goal include speedy data gathering, the imagining of epidemic data, the interruption of wide spread risk, the following of complete cases, and the monitoring of preventative levels for COVID-19. In this study, the authors analyse and forecast COVID-19 inside a health monitoring system. The framework makes use of big data analytics and the IoT. With the help of big data branches, we achieve evocative, analytical, prognostic, and inflexible analyses of a novel illness data set that focuses on an extensive assortment of pandemic symptoms. The fundamental contribution of our work is participating giant data and Internet of Things to assess and forecast a rare disease. The deep learning & machine learning models may be used to identify and forecast the epidemic, which would be helpful to medical staff. Pandemic predictions are made using a range of ML methods. Additionally, GNB excels in comparison to other solutions, as seen by its accuracy rate of 81.5%.Item Classification of multiforme glioblastoma using deep learning technique(UMT, Lahore, 2022) Hamza RamzanThe researchers respective to the domain of Artificial Intelligence(AI), Machine Learning(ML) , Deep Learning(DL) , Data Science(DS) and Reinforcement Learning(RL) are creating a huge positive change in the world. These computing domains are leading humans in a new scientific age of development of autonomous systems that can be equipped at various area including robotics, autonomous vehicles and smart logistics, intelligent homes and industries, medical and health sciences, intrusion and fraud detection and entertainment industries too. Cancer is a collective term that includes a lot of sub-diseases that are increasing rapidly worldwide, Cancers are highly deadly, it has increased the human mortality rate to some extent and requires early and correct diagnosis so the treatment can be started soon so the patient can be saved. Symptoms of the different types of cancer are quite relevant and analogous that they are often fault diagnosed with some other types while the symptoms don’t arise in the initial stages. Brain tumors are one of the most prominent and incident types of cancer found in humans and also consists of subtypes as well. The proposed methodology presented is developed to tackle this issue, a deep learning based system is introduced that will be helpful for the early diagnosis of Multiforme glioblastoma and its subtypes. The Dataset was taken from the Cancer Imaging Archive and consist of 4 classes 3 classes represents 3 types of glioblastoma and 1 class represents no tumor. It is a multi-class classification system built using a VGG16 model equipped with Deep Learning technique the model was successful to produce a training accuracy of 99.43% while an overall validation accuracy of 99.07% with a 99% score for precision, recall and F1-score respectively.Item Metaheuristic method for energy management of microgrid(UMT, Lahore, 2022) MARYAM AYAZThis thesis proposes the use of metaheuristic optimization techniques to minimize the operating cost of Micro-Grid Energy Management System (MG EMS). The primary contribution of this research is the implementation of these techniques to provide an efficient and cost-effective solution for managing the energy in micro-grids. The optimization strategy aims to achieve the global optimum solution by considering time of use (TOU) and renewable energy sources. The goal is to decrease expenses related to energy and improve the effectiveness of energy usage. The proposed optimization strategy is a crucial aspect of this research as it seeks to improve the economic viability and sustainability of micro-grids. By minimizing the operating cost, the system's overall performance can be enhanced, ensuring reliable and durable operation. The implementation of metaheuristic optimization techniques will also help in overcoming the limitations of conventional optimization methods by enabling the system to search for solutions more efficiently and effectively. I utilized the Symbiotic Organisms Search Algorithm (SOS) for my thesis to minimize the operating cost of Micro-Grid Energy Management System (MG EMS). Compared to different well-established optimization algorithms, SOS outperformed in optimizing the operating cost of MG EMS. The results of this research can be used to inform policymakers and energy management professionals about the potential benefits of using metaheuristic optimization techniques in micro-grids. The proposed approach can be used as a reference for future research in the field of energy management systems. It has the potential to contribute to the development of more cost-effective and efficient energy management models for micro-grids, thus enhancing the overall sustainability of the energy sectorItem Machine learning methods for cyber attack detection in IoT(UMT, Lahore, 2022) Habiba HabibItem Plebiscite prediction using social sentiment analysis(UMT, Lahore, 2022) IQRA MUBARIKThe rise of social media, Now the modern customers now have a powerful voice. Businesses (or analogous organizations) need to know the polarity of these views in order to have a deeper understanding of user orientation and make smarter decisions. A good example of where this might be beneficial is in politics, where a candidate's or party's fortunes may rise or fall depending on how accurately they predict voter opinion. Some have discovered that analyzing social media data for sentiment is a helpful way to track user preferences and habits. Both Naive Bayes and Support Vector Machines (SVM) are supervised learning algorithms that are often used for text categorization, and both need a training data set in order to perform sentiment analysis. Whether or not these algorithms succeed depends heavily on the quantity and quality (features and contextual relevance) of the tagged training data. Because of a dearth of training data, most applications rely on cross-domain sentiment analysis, which fails to capture details unique to the target data. It reduces the accuracy of text classification as a whole. Here, we offer a two-stage approach that may be used to produce training data from the mined Twitter data without compromising features or contextual relevance. Next, we provide a machine-learning model for election prediction that is scalable and based on our two-stage procedure.Item Implementation and analysis of internationally recognized network security standards for smart energy meters(UMT, Lahore, 2022) ANAM SHAHZADINetwork is always important component in any organization for their operational tasks like network control communication and maintain flow of data and technological requirements. On the other hand, the execution of security standards in network has been done manually in the past. For many organizations the other aspects of networks and standards execution was manual. The major issues of these traditional practices were that there were time consuming, costly, and complicated and have no standardization. But the today study in networking filed require latest technology for the security standards implementation. When we talk about the automation in implementation of standards it will be revolution in this filed and gives new directions for future. Hence, this work presents the idea of implementing the standards in automated simulated environment for creating security module and integrating this with security standards of IEEE and Cisco best practices.Item Trends of blockchain technology in a variety of life domains(UMT, Lahore, 2022) SYED FARHAN RAZA NAQVIThe research and business communities have recently begun to place a greater emphasis on blockchain technologies because of the potential benefits that they could bring to a wide range of industries. This is as a result of their practical capabilities in resolving a multitude of issues that are presently preventing further advances in a variety of industrial domains. Some examples of these issues include the establishment of automated and efficient supply chain processes, the enhancement of transparency throughout the entirety of the value chain, and the recording and sharing of transactional data in a secure environment. Blockchain technology provides an efficient solution to these problems by utilizing transactional ledgers that are decentralized, shared, encrypted, and permissioned. The use of blockchain systems and their potential for implementation in a variety of circumstances enables a broad range of industrial purposes by escalating safety and security, improving efficiency and transparency, and controlling costs. This paper analyses and discusses a wide range of different commercial application fields where the use of blockchain has been recommended. This research looks into the possibilities, advantages, and hurdles that can emerge from utilizing blockchain in a wide range of commercial applications. Furthermore, the paper attempts to identify the requirements for the smooth completion of blockchain across a broad range of industries. The study indicates that distributed ledger has a number of possible uses across a broad spectrum of industries; however, there remain a few barriers that must be resolved in order to maximize the use of this technology.Item Design of a cipher using cubic dynamic substitution box(UMT, Lahore, 2022) FARRAH FAREEDNowadays, many organizations, companies, and an individual person share their data online through unreliable mediums ignoring that there are many hackers sitting outside to hack their data and misuse it. A cipher in cryptography is a set of predetermined procedures that may be used as a technique to achieve encryption and decryption. Cryptosystems are employed to demonstrate the security of private communication and information networks. Various permutation and substitution techniques are employed for this purpose. S-Box is in charge of the substitution procedure. Modern ciphers employ dynamic, key-dependent S-Box. In this work, we suggested a way for creating dynamic S-Boxes with strong cryptographic features. Techniques, such as bijectivity, nonlinearity, strict avalanche criteria, and linear probability, are used to assess the strength of the S-Box. After the creation of a dynamic S-Box, it is used in a novel cipher designed in this research thesis. Novel cipher consists of three phases such as key generation, encryption, and decryption. A different S-Box is generated using values taken from the key and employed in a cipher round to make things more difficult for the attackers.