Department wise Listing | NUML Online Research Repository
List of Content
Back to Listing
Title Abstract Action(s)
A STUDY ON ISSUES OF ROMAN URDU TEXTING BY OLDER USERS As the development and utilization of technology is increasing rapidly. It identifies different experiences for older users. Most of the time, older users face issues in utilization of new interface. According to the United Nation world population report, the older people will become the main and largest part of the population in the coming era. With the increasing population of older people, Pakistan will also become an aging society. Older users are the late adopters of smart mobile phones technology. The older users in Pakistan are ready to use digital communication ways but they often face issues when they do Roman Urdu texting. Different factors are affecting their approval of Roman Urdu texting communication. The available smart mobile phones’ texting applications do not fulfill the needs of older users because older users are minority of digital system and minorities are poorly represented in digital system. To make the human computer interaction community aware, there is need to take the older users into account. There is also need to take notice about very common feature of mobile phones which is texting. Texting is an important and mostly used feature of mobile phones. As compared to the other digital communication mediums, texting is adopted by large part of the population of the world. It is observed that older users do no use texting applications. They often face issues when they interact with texting applications. Thus, the aim of this research study is the investigation of Roman Urdu texting issues faced by older users of Pakistan. Two phase methodology is used to inquire the texting issues. The first phase of methodology is literature review, to the investigate of texting issues faced by older users worldwide. The second phase of methodology is survey, to the investigation of Roman Urdu texting issues faced by older users of Pakistan. The results of this study are the Roman Urdu texting issues faced by older users and guidelines. Guidelines are proposed in order to minimize the Roman Urdu texting issues. The results of this study may be helpful in the design of texting applications for the older users of Pakistan.
IMPACT OF PERFORMANCE REWARDS ON EMPLOYEE TURNOVER IN THE SOFTWARE INDUSTRY OF PAKISTAN Impact of Performance Rewards on Employee Turnover in the Software Industry of Pakistan Employee turnover crisis has its negative outcomes spreading almost every field. Software industry has been facing it for quite some years now. Many organizations take necessary measures to ensure employee retention, performance rewards are one of the measures. However, not all performance rewards tend to be as motivating as considered. The study emphasizes the use of performance rewards in the Software Industry of Pakistan. Moreover, it is being intended to study and identify the key rewards that give a real boost to the motivation of employees hence lowering the chances of employee turnover. The study uses survey (questionnaires) methodology to identify the types of awards being awarded to the employees, the psychological impact of performance rewards and the performance reward that can help the best in minimizing employees’ turnover. The study has been conducted on the managers and employees having technical roles in the Pakistani software industries. The research has indicated that performance rewards, especially the monetary performance rewards play a vital role in minimizing the employees’ turnover. In some cases, company swag, verbal appreciation and promotions also showed driving the inclination of an employee towards working in the organization. Furthermore, it has also come into knowledge that factors like reward frequency and amount of monetary reward value have also got a significant impact on the decision of an employee on serving the organization or leaving it. The findings of the study if implemented in true letter and spirit will help the managers determine the best reward schemes and it should also contribute in the Business Analysis Body of Knowledge (BABOK).
TAXONOMY OF SUSTAINABILITY INDICATORS FOR SOFTWARE DEVELOPMENT The rise in environmental and social issues has led to increase the awareness of the sustainability of human activities. The software development industry is no exception, and there has been a recent surge in the number of studies on sustainable software development (SSD). However, still, there is not enough understanding of what constitutes SSD and how it can be achieved. The problem being addressed in this study is the lack of a comprehensive taxonomy of sustainability indicators for software development. Systematic literature review was performed to identify existing SSD indicators. The results were analyzed and classified into various dimensions of sustainability. The taxonomy can be used by the practitioners to select appropriate indicators for sustainable development of their software projects. A quantitative approach is adopted by the researcher as the following research design aids in collecting primary data that has the ability to afford results with high precisions along with appropriate statistics. The method of survey questionnaire has been used in the current research in which the information has been collected from 80 respondents. The focus of the present research is on the primary quantitative method. Therefore, the data has been analysed using Statisical Package for Social Sciences (SPSS) software through frequency analysis, regression and correlation analysis. It has been found that the heightened awareness of the degree to which human activities can be maintained has come about as a direct consequence of the rise of environmental and social concerns. The industry of software development is not an exception, and there has recently been an increase in the number of studies that focus on environmentally responsible practices in the software development process. It is recommended that significant value and consideration should be given to the paradigm of green software system as it has mainly assist the corporations to develop eco-friendly system, which is cost-effective as well.
A STUDY ON USABILITY FACTORS TO RETAIN CUSTOMERS Customer retention refers to activities and actions taken by organizations in order to retain customers. The key factor of growth and long term profits of any organization is customer retention. Different studies have been conducted to understand and identify possible factors that contribute in customer retention. User interface of mobile applications potentially plays an important role in retaining customers especially in context of mobile applications. There is a need to further explore the role of user interface in customer retention. The main aim of this research is to study usability factors to retain customers in the context online ride hailing services. Firstly, literature review is done to find out the possible usability factors that could led to customer retention. Mostly mentioned usability factors by the researchers are ease of use, effectiveness, efficiency, usability (interaction with interface) and user satisfaction. Secondly, these factors are used as a base to develop questionnaire for the survey. A survey is conducted to see whether there is a correlation between usability factors and customer retention, particularly in the context of ride-hailing services. The sample size for the survey is 446. After doing statistical analysis by using Pearson’s Correlation method it is proved that there is positive relationship between customer retention and usability factors in case of ride hailing services. Effectiveness is one of the factor with highest frequency in literature review and it also got highest value 0.7145 that is greater than critical value 0.09 so it proves that it is an important factor that should be taken into consideration while developing any application in order to retain customers. Based on the statistical analysis all factors significantly contribute towards user retention. Ease of use is the factor having least significant value 0.5363 as it is closely related to the critical value so this factor can be said as showing signs of least significant factor as compared to other factors.
FRAMEWORK FOR MINIMIZING THE IMPACT OF UNCERTAINTY IN SOFTWARE PROJECTS DURING COVID 19 PANDEMIC Framework For Minimizing the Impact of Challenges in Software Projects During Covid-19 Pandemic The novel Corona Virus was originated in Wuhan, China, in December 2019 and was declared a global pandemic in March 2020 by the WHO (World Health Organization). Covid-19 has been disastrous; it has influenced every aspect of life around the globe, with a death toll of millions till date. Global pandemic has changed how every human, institution, and organization works. The fatal virus and its spread also targeted the software industry. This research analyzes how the COVID-19 pandemic has affected software projects and some of the challenges faced by the teams. A mixed research methodology has been used, including the systematic literature review. Further, to add empirical proof, information has been gathered from different software houses by conducting an online questionnaire-based survey. One-on-one interviews are also conducted to validate the survey results. The research has identified some of the challenges and proposed valuable mitigation techniques by keeping the local working culture of Pakistan in mind. Based on the gathered information, a conceptual framework has been proposed and initially validated by experts in the respective field
PREDICTION MODEL FOR THE REDUCTION OF YOUNG DRUG ABUSERS The problem of drug addiction is increasing day by day at alarming levels. The understanding of addictive disorders and psychiatric pathologies has become easier through new computational technologies and techniques. Collection and comparison of data has become more efficient through the usage of new emerging AI trends. Technique of digital phenotyping paves the way for capturing characteristics of different psychiatric disorders in patients. Likewise, machine learning is helping the doctors in the classification of patients based on different patterns detected through data. Almost 40,000 people are becoming drug addicts in Pakistan annually. Drug addiction problem is caused due to many reasons like peer influence, curiosity or family disturbances. This research focuses on those drug addicts who have stepped in this social evil due to some family issues. The best possible solution for controlling this social evil is to bring awareness among the parents about the effects of their behaviors on the mental and physical health of the child. In order to do that predictive analysis was applied to forecast the upcoming trends and events in drug addiction due to family disturbances. First systematic literature review was conducted for deducing the major family factors effecting the health of child from extensive literature. Six family factors were inferred parent child activities, family structure, parent child communication, parents involved in drugs, parent monitoring and supervision, and strategies for family management. After the SLR, survey was conducted from drug addicts in order to gather data for predictive analysis. During the survey age of the patients was limited to 13 till 25. Total 3528 patients have been selected for the study. However, twin cities have been targeted for the data collection purpose. After the collection, data was wrangled and labeled properly and three classification models were applied Naïve Bayes, Decision Tree, and Random Forest. Decision Tree had the maximum accuracy percentage of 96%. After that upcoming trends were depicted for the six factors. The current values of family factors are 747, 430, 1018, 296, 1497, and 437 respectively. The predicted values are 4455, 2321, 3895, 5353, 25417, and 9098 respectively. By reviewing these values it’s evident that government needs to take quick actions against this social evil and parents need to be acknowledged about the impact of their actions on children.
EXPLORING THE MOTIVES AND ADDICTIVE BEHAVIORS OF ESPORTS PLAYERS IN PAKISTAN With changing lifestyles, outdoor games have been replaced by esports games, which are rapidly gaining popularity among young players in Pakistan. These esports games are a kind of competition in which players compete against one another, either in teams or individually, following a set of rules. However, despite the popularity of esports games, there is a lack of research on players' motives and behaviors symptoms of players in Pakistan. Hence, the objective of this study is to determine how prevalent esports games are among young players in Pakistan. To fulfill this objective, a survey methodology was used. Survey questions on demographics, esports motives, and addictive behavioral symptoms were asked by participants using convenience and snowball sampling techniques. A total of 420 esports players participated in the survey. For the analysis of the data, factor analysis was performed to examine the variables of esports game consumption. The results reveal six motives to play esports games: violence, entertainment, coupling, fantasy, psychological benefits, and skill enhancement. As for the behavioral symptoms, salience, withdrawal, mood modification, and loss of control emerged as the four main symptoms. Additionally, survey results reveal that esports games do carry positive aspects of enhancing social skills and interactions among the players while helping them exhibit behaviors and emotions that are not coherent with mental disorders.
EEG-BASED EMOTION ANALYSIS USING TEXT STIMULI Emotions are fundamental to our daily lives, impacting our behavior, thoughts, and feelings. Accurately classifying emotions is of utmost importance in various fields, including psychology, psychiatry, and neuroscience, as it can aid in the development of effective diagnostic tools, and human-computer interaction systems. In this study, the aim was to classify emotions using Electroencephalogram (EEG) data, which provides a non-invasive and objective measure of brain activity. Data was collected from 25 participants using a 128-channel device and text stimuli. To achieve this goal, a comprehensive approach was adopted that integrates multiple feature extraction techniques, pre-processing techniques, and a Support Vector Machine (SVM) classifier. Four feature extraction techniques, Convolutional Neural Network (CNN), Wavelet Transform (WT), Power Spectral Density (PSD), and Raw data, are used to extract features from pre-processed EEG signals. The pre-processing techniques involved down sampling, re-referencing, and filtering the EEG signals to eliminate noise and artifacts. The Monte Carlo approach is applied to randomly selecting training and testing samples to ensure the reliability and validity of results of this study. Study focused on classifying emotions as positive and negative. T-tests were used to identify the most relevant features that contributed to the classification of emotions. The results show that the combination of CNN and SVM yields the highest average accuracy rate of 80%, followed by WT with 75%, PSD with 72%, and raw data with 65%. This suggests that the use of CNN, WT, and PSD as feature extraction techniques in combination with SVM as a classifier can significantly improve the classification of emotions based on EEG data. The proposed approach has significant implications for the development of more accurate and efficient methods for classifying emotions in various fields. For instance, in the field of human-computer interaction, accurate emotion recognition can be used to develop personalized interfaces that respond to the user's emotional state. In clinical settings, the accurate classification of emotions can aid in the diagnosis of psychiatric disorders and inform treatment strategies. Our findings highlight the potential of EEG data as a valuable source of information for emotion classification, with important applications in various domains.
Detection And Classification Of Brain Tumor Using Deep Learning A brain tumor is a potentially fatal condition caused by uncontrolled brain cell development that affects human brain cells and the neurological system. Brain tumors are among the leading causes of death worldwide. Therefore, for a patient to receive appropriate medication, a precise and early diagnosis of a brain tumor is essential. It also helps to avoid time-consuming and painful medical procedures. Manual brain tumor identification and treatment takes a long time, is complicated, and contains a human error. As a result, accurate automatic techniques are needed for the segmentation and classification of tumors. Machine learning techniques are employed for early detection and enhanced results. Precise tumor segmentation and classification are important in radio surgical planning and evaluating tumor treatment efficacy. The purpose of this research is to develop a deep learning-based system for segmenting and classifying brain tumors. In this study, a 3D U-Net model is used for the segmentation of MRI images, followed by 3D CNN for the classification of segmented images. BraTs 2019 datasets are used for training and testing the model. This model gets the accuracy of 96%.
DETECTION OF FAKE NEWS USING NATURAL LANGUAGE PROCESSING AND STATISTICAL TECHNIQUES Detection of Fake News using Natural Language Processing and Statistical Techniques The rapid growth of fake news in online media and social platforms has become a major concern in recent years. To address this problem, we propose a system that integrates Statistical techniques, Machine Learning (ML), and Natural Language Processing (NLP) methods to identify fake news. We extract features using the Term Frequency- Inverse Document Frequency (TF-IDF) method and evaluate the classifier’s performance using metrics such as accuracy and precision. We have used three different ML algorithms for classification, namely Naive Bayes (NB), Support Vector Machine (SVM), and Logistic Regression (LR). Our results show that LR with Maximum Likelihood Estimation (MLE) outperforms other classifiers, achieving an accuracy of 95%. Our study includes the use of NLP methods for feature extraction, which improves the accuracy of fake news detection. In addition, we demonstrate that LR with MLE is an effective approach for identifying fake news by reducing the complexity and dimension of features, which can help to prevent the spread of fake news. The study provides a valuable contribution to the field of fake news detection and highlights the importance of integrating NLP methods with Statistical and ML techniques to improve the effectiveness of fake news detection.
Capacity optimization in smart grid communication network using Non orthogonal multiple access In contrast to the conventional electric grid, a smart grid (SG) is an innovative and modernized electric grid system. Specifically, the electric grid is upgraded by advances in information and communication technologies, where terabytes of data are generated at SG devices that must be transmitted. The anticipated size of data originates from various applications such as metering, automation, monitoring, and firmware updates. Therefore, the smart grid communication network (SGCN) requires a communication system that is spectrum efficient and does not overburden already scarce channel resources. Meanwhile, non-orthogonal multiple access (NOMA) has been envisioned to support the demands of latency, throughput, and fairness of future wireless networks. Mainly, NOMA has been studied for the objectives of capacity optimization, fairness, and energy efficiency for various communication situations. The investigation for NOMA for domains of cellular communication underwater as well as unmanned aerial vehicle (UAV) communication has demonstrated, that it has a clear advantage over conventional Orthogonal Multiple Access systems. Though, NOMA's applicability to SGCN has not undergone examination yet, and further investigation is needed. This thesis will deal with the design issues related to user pairing and power allocation for effective implementation of NOMA in the context of SGCN.
Classification of Cyber attacks on IoT using Machine Learning Internet of Things (IoT) devices are growing rapidly, which are raising concerns regarding data security and privacy. With the increase in volume of data generated by IoT devices, secure transmission and processing has become essential. Protecting IoT infrastructure and devices from cyber-attacks is an active research area now-a-days. The detection and classification of attacks on IoT require dynamic and automatic techniques, based on Machine Learning (ML). This thesis focuses on the classification of cyber-attacks on IoT using ML. It explores the potential of ensemble learning to enhance the accuracy of ML. Two different publicly available datasets are used in this thesis. These datasets are UNSW NB 15 and NSL KDD. The ensemble learning is implemented by combining the strengths of Extreme Gradient Boosting (XGBoost) and Deep Neural Network (DNN). The proposed ensemble learner is named as X-DNN. The performance evaluation of X-DNN is carried out using, accuracy, precision, recall, and F1 scores. The experimental results show that X-DNN demonstrates an accuracy of 85.36% in classifying cyber-attacks on UNSW NB 15 dataset whereas it achieves accuracy of 99.81% on NSL KDD dataset. X-DNN outperforms state of the art methods such as Hierarchical Clustering Decision Tree Twin Support Vector Machine, Gated Recurrent Unit, and Deep Neural Networks. The proposed method offers relatively better performance which highlights its significance in classifying cyber-security attacks on IoT.
ANALYSIS OF FACTORS AFFECTING WIND TURBINE ENERGY OUTPUT USING MACHINE LEARNING
ANALYSIS OF FACTORS AFFECTING WIND TURBINE ENERGY OUTPUT USING MACHINE LEARNING Electrical energy generated by wind turbines is stochastic in nature due to its dependency on various factors. Such randomness raises barriers in adjusting the energy stocks of the power systems according to need. Multiple approaches have been proposed to predict the energy output of wind turbines and to meet the corresponding energy demands. This thesis investigates variables (also called factors or features) that affect the wind turbine’s output. The energy obtained from turbines varies as it relies on factors. Some of the important factors or features are turbine blade area, wind speed, temperature, air density, humidity, tower height, the angular position of the blades, pressure, etc. All such features are required to be investigated, analyzed, and evaluated with the help of state-of-the-art Machine Learning (ML) models to identify their importance and significance in predicting the wind turbine output. ML techniques used are Categorical Boosting (CatBoost), Extreme Gradient Boosting (XGBoost), Decision Tree (DT), Random Forest (RF), Gradient Boosting Regression (GBR), Light Gradient Boosting Model (LGBM), Extra Tree and Adaptive Boosting (AdaBoost). Evaluation of the ML methods and analysis of the factors are carried out on three different latest and publically available datasets. The experimental results show that CatBoost compared to all other methods demonstrates the lowest Mean Absolute Error (MAE) and Root Mean Squared Error (RMSE) in predicting the energy output. The wind speed is identified as the most significant factor in predicting the energy output by all of the ML methods. Additionally, a new method is proposed which ensembles CatBoost, RF and AdaBoost methods. The proposed method is named as X-CRA, in which the prediction s of CatBoost, RF and AdaBoost are fed into XGBoost through a stacking approach and final energy output is obtained. The experimental results show that X-CRA outperforms CatBoost and all other ML methods in predicting the wind energy output.