Dr. Muhammad Farhan | COMSATS Institute of Information Technology sahiwal (original) (raw)

Papers by Dr. Muhammad Farhan

IoT based smart interaction framework for elearning

Proceedings of the 33rd Annual ACM Symposium on Applied Computing, 2018

Multimedia content boosts the learning trends. This Multimedia and Agents based Question Answerin... more Multimedia content boosts the learning trends. This Multimedia and Agents based Question Answering System (MAQAS) is based on Internet of Things in eLearning paradigm. It is aimed to give help especially to the students in finding the more relevant and accurate answers to their questions. This system is for both synchronous and asynchronous communications. The multimedia-based IoT-centric environment is suitable to enhance the effectiveness of the delivery of learning contents. This creates a flexible eLearning paradigm for the teachers and students. The proposed eLearning model uses sensors to detect the student location, temperature, and mobile camera to identify the student's presence at a suitable place, inappropriate atmosphere and his activeness in a learning environment, respectively. Three agents are playing vital roles in making the smart decision about student-teacher interactions. Experimentation is performed and the initial results are drawn. Accuracy value of MAQAS is compared with the well-known existing QA systems; LIVE QA TRACK, QUORA, YODA QA LIVE, AND ASKMSR-QA concerning three types of WH Questions. The overall performance of the proposed systems is proved to be better compared to other competing systems. It was also concluded that eLearning with multimedia settles the goodness of an IoT-based solution. It also enhances the QA to Multimedia and Agent-based Question Answering System.

International Journal of Distributed Sensor Networks, 2019

It has been seen that most of the accidents occur due to driver’s fatigue. Drowsiness is a state ... more It has been seen that most of the accidents occur due to driver’s fatigue. Drowsiness is a state of mind before the driver falls asleep, which means the driver could not accomplish his actions, such as vehicular braking, controlling vehicular motion, properly. We have built an Internet of things–based medical application to analyze driver’s drowsiness. An architecture has been proposed and a simulation of that scenario in NS3 WSN simulation tool has been done. This simulation shows that the ratio of accidents can be majorly reduced. When drowsiness of drivers is captured, a message alert is delivered to all other drivers of the vehicles that are near to the sleeping driver; for this, different sensor nodes are used. Another unique feature of the sensor network used here is the collaborative effect of sensor nodes. So for measurement and analysis of applications on Google Play, a dataset of the medical applications category was scraped. The scraping was done with 550 applications of ...

Recently in many works, it has been identified that large-scale multiple-input-multiple-output (L... more Recently in many works, it has been identified that large-scale multiple-input-multiple-output (LS-MIMO) is one of the key technologies for achieving extraordinary gains of energy and spectral efficiencies in wireless communication systems. This technology employs low cost hardware which enlightens its significance. The efficiency of single-cell or multi-cell LS-MIMO systems is affected by various factors such as hardware impairments, energy consumption, bandwidth allocation and outdated channel. In this paper LS-MIMO systems have been analyzed with the effects of channel aging or often called outdated channel. New expression for minimum-mean-square error (MMSE) estimate has been derived for LS-MIMO system in downlink (DL). We use three precoding schemes and results show that regularized zero-forcing (RZF) gives better performance than zero-forcing (ZF) and maximum ratio combining (MRT). We also show that how Doppler's shift in combination of outdated channel affects the performance of the system with all the above mentioned precoding schemes.

The IPv4 address architecture has been declared ended finally due to the fast growth of the Inter... more The IPv4 address architecture has been declared ended finally due to the fast growth of the Internet of Things (IoT). IPv6 is becoming a next-generation communication protocol and provides all the requirements that the industry needs. A smart home is an emerging technological revolution in which IoT-enabled smart physical objects such as smart TVs, smart refrigerators, smart locks, etc. are linked to the Internet to make human life more comfortable. There are several resource-constrained smart devices interconnecting with 6LoWPAN to control the smart home remotely. The communication channels used by cellular communication are vulnerable and increase security threats especially related to authentication. A reliable and portable remote authentication method is critical for ensuring safe communication in the next-generation smart home environment. Recently, many authentication schemes have been proposed but adopt complex mathematical techniques or protocols that are viewed as heavyweights in the context of computation and communication costs. This research proposes a lightweight and reliable remote authentication mechanism for the next generation of IoT-based smart homes. Informal and formal security assessments through the AVISPA tool determine the robustness of our proposed scheme. Moreover, we implemented our authentication scheme on a Linux-based client-server network model by using Android programming. In addition, we compared our proposed scheme with existing schemes based on computation and communication costs. Results show that our proposed mechanism reduced computation costs by up to 54.03 % and reduced communication costs by up to 25.28 % related to existing schemes. So, our proposed scheme is better, more secure, and most suitable for smart home ecosystems.

The stock market index value is a useful tool for investors, public businesses, and governments t... more The stock market index value is a useful tool for investors, public businesses, and governments to invest money while considering the potential for profit and danger of loss. In financial data analysis, the prediction is applied widely to enhance the accuracy of forecasting individual stock indexes and correlation to other indices of other stock market companies. This paper analyses and forecasts time series over the specific span of days aiming at the Karachi stock market. There are different attributes of datasets like Symbol, Date, Open, High, Low, Close and Volume; each attribute has a significant description that plays an essential role in the machine learning analysis. Dicky-fuller statistics have been applied to convert the data into stationary time series. It is used to analyze the stock statistics behaviour, extract trends and seasonality. Time difference lag is used to smooth the data. The data is decomposed to remove trends and seasonality after the transformation in the data. Model diagnostics are applied to analyze the fitness of the model on the data. Time series data prediction is performed by applying the Seasonal ARIMA predictor in conjunction with Autocorrelation Function (ACF) and Partial Autocorrelation Function (PACF).

The forecast of frost occurrence requires complex decision analysis that uses conditional probabi... more The forecast of frost occurrence requires complex decision analysis that uses conditional probabilities. Due to frost events, the production of crops and flowers gets reduced, and we must predict this event to minimize the damages. If the frost prediction results are accurate, then the damage caused by frost can be reduced. In this paper, an ensemble learning approach is used to detect frost events with Convolutional Neural Network (CNN). We have used this to get more efficient and accurate results. Frost events need to be predicted earlier so that the farmer can take on-time precautionary measures. So, for measurement and analysis of Google Play, we have scrapped a dataset of the Agricultural category from different genres and collected the top 550 application of each category of Agricultural applications with 70 attributes for each category. The prediction of frost events prior few days of an actual frost event with an accuracy of 98.86%.

The world today has become a competitive place, whether it is in sports, entertainment, or in the... more The world today has become a competitive place, whether it is in sports, entertainment, or in the job market. Users need to allocate an extraordinary amount of time for their routine works with little distraction. It is a truism that time is in all works of life or all aspects, in all aspects of life. They found out that people generally spend one hour daily in front of the mirror. This paper introduces the Interactive Multimedia-futuristic Live Mirror by using the concept of Artificial Intelligence for commercial use as well as for the smart home environment. The data collection for this work is done by using the cloud with Arduino UNO. In one smart city, several live mirrors were attached in different smart homes and industries in different smart homes and industries as they all invariably collect data from the cloud and stay updated. Additional, information can be collected from home appliances such as deep freezer, air conditioner, and washing machine. The RFID, with a Master-Slave architecture, is used to construct a smart-home system, displaying all information on the Live Mirror. Our smart mirror is equipped with several digital devices such as speakers, microphone and Arduino, which enable it to provide essential standard services such as suggestion, calorie monitoring, activities, latest health habit updates, and current weather. Therefore, it makes it a new smart product with growing global recognition in recent years.

Reliability of data is severely affected due to high attenuation, high noise, limitation of the a... more Reliability of data is severely affected due to high attenuation, high noise, limitation of the acoustic medium, and dynamic topology. The limitation of the acoustic medium, attenuation, and noise, much work was conducted by using forward error correction process, retransmission, attenuation model and Signal to Noise Ratio (SNR) to make data reliable. Many routing protocols are introduced to increase the lifetime of the network. However, there are also many improvements which need to be done to enhance the lifetime of the network. In this research work, we introduced a Congruent Gravity Value (CGV) routing protocol to enhance the lifetime of the network. Two types of packets will be transmitted in order to find out the relay node. One type of packet is hello packet, which will contain the information of source node and sink. The second type of packet is a response packet, which contains the information of the nearest neighbor node. Hello, the packet will always transmit by source node while the response packet will be transmitted either by the nearest neighbor node or by source node based upon condition set. That condition will save energy and enhance the network lifetime whenever it is feasible. Gravity function plays a preeminent role in the selection of the relay node. In order to tackle the Congruent gravity value, the cluster head selection technique is used. Network Simulator has been used to evaluate the performance of CGV, and its results are compared with ER2PR, VBF, and UFCA. These protocols utilized a large area in order to select a relay node. To overcome this problem, CGV used just those area which is best for the selection of the relay node and response packet transmission area is also minimize whenever it is feasible by the condition.

This research aims to an electronic assessment (e-assessment) of students' replies in response to... more This research aims to an electronic assessment (e-assessment) of students' replies in response to the standard answer of teacher's question to automate the assessment by WordNet semantic similarity. For this purpose, a new methodology for Semantic Similarity through WordNet Semantic Similarity Techniques (SS-WSST) has been proposed to calculate semantic similarity among teacher' query and student's reply. In the pilot study-1 42 words' pairs extracted from 8 students' replies, which marked by semantic similarity measures and compared with manually assigned teacher's marks. The teacher is provided with 4 bins of the mark while our designed methodology provided an exact measure of marks. Secondly, the source codes plagiarism in students' assignments provide smart e-assessment. The WordNet semantic similarity techniques are used to investigate source code plagiarism in binary search and stack data structures programmed in C++, Java, C# respectively.

There are millions of applications uploaded by the developers on the daily basis. Without any che... more There are millions of applications uploaded by the developers on the daily basis. Without any check and balance millions of users download these applications. Theses duplicated applications damage the users trust on Google play store and can grab the confidential information of user. There is no more information provided by developers on the front end of the application that can define the legitimacy of the application. In this paper, by using a Google-play-scraper build a Google play store dataset with all categories of games. Scraping at least 550 applications of each category of games in free and respectively in paid applications by using Google play scraper, cumulatively scrape the 3600 paid applications and 10k free applications of all categories in games. The categories of these games' applications use respectively are Word,

Acquiring sensitive information from the user in some malicious web pages which looks like the le... more Acquiring sensitive information from the user in some malicious web pages which looks like the legitimate webpage and they do a kind of criminal activity that is known as phishing in the electronic world. An attacker can use this kind of phishing or fraud by using such websites, which is a severe risk to web users for their personal and confidential information. So, in the field of e-banking and e-commerce, this act makes a threat for all webpage users. In this paper mainly discerning the different features of legitimate, suspicious and phishing websites. These features are fed to the machine learning algorithms which are built-in WEKA are used for comparison and to check the accuracy of the algorithm. Algorithms used in this comparison are J48, Naïve Bayes, random forest and Logistic Model Tree (LMT) are used and them accurately to predict the website legitimacy is calculated. Also, the best algorithm among different algorithms can be selected. In this paper, we will compare the results in the two ways. Firstly, we find the best algorithm by using the comparison of the different attributes like Correctly Classified Instances, Incorrectly Classified Instances, Mean absolute error and kappa statistics. Secondly, the accuracy of these algorithms will analyze with different parameters like TP Rate, FP Rate, Precision, Recall, F-Measure, MCC, ROC Area and PRC Area that is visualized in the bar chart. The selected algorithm makes the website analyzing process automated. Before making payment on any e-commerce website, this prediction model can be used for determining the legitimacy of that website.

In a smart city, the use of wireless sensor network is an emerging technology, and it provides ma... more In a smart city, the use of wireless sensor network is an emerging technology, and it provides many benefits in terms of communication, energy, and cost. We have focused on energy efficient smart lighting system. In our proposed model we have divided street lights energy usage into three categories: low, moderate and high. The street light energy usage is low in daylight, moderate when average traffic on roads and high when heavy traffic is on roads. Street lights turn on when a vehicle enters in a passage after sensing its entry within a region. We have deployed our system on street lights with the facility of solar panel so that our devices have more life as in daylight it gets charged, and at night it turned on based on the presence of a vehicle.

The Software plagiarism, which arises the problem of software piracy is a growing major concern n... more The Software plagiarism, which arises the problem of software piracy is a growing major concern nowadays. It is a serious risk to the software industry that gives huge economic damages every year. The customers may develop a modified version of the original software in other types of programming languages. Furthermore, the plagiarism detection in different types of source codes is a challenging task because each source code may have specific syntax rules. In this paper, we proposed a methodology for software plagiarism detection in multiprogramming languages based on machine learning approaches. The Principal Component Analysis (PCA) is applied for features extraction from source codes without losing the actual information. It extracts features by factor analysis and converts the dataset into normalized linear principal components which are further useful for predictions analysis. Then, the multinomial logistic regression model (MLR) is applied to these components to classify the source codes documents based on predictions. It gives the generalization of logistic regression to handle multiclass problems. Further, the predictors' performance in MLR is evaluated by 2 tailed z test. To apply the experiment, the dataset is collected in five different and popular languages, ie, C, C++, Java, C#, and Python. Each programming language taken in two different case studies, ie, binary search and Stack.

With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threat... more With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threats have increased. Identifying malicious attacks in IoT requires advanced techniques tailored to this ecosystem. Existing algorithms have limited effectiveness in detecting obfuscated IoT malware. This study proposes the elucidating cybersecurity-promulgated malware taxonomy (ECMT) framework, combining memory analysis and ensemble machine learning (ML), to enhance IoT malware categorization. ECMT integrates support vector classification, quadratic discriminant analysis, and AdaBoost on forensic artifacts from memory dumps to improve detection across families, such as ransomware, spyware, and trojans. ECMT can enable intrusion prevention, information protection, and cybercrime deterrence in IoT environments. Experiments on a balanced data set indicate AdaBoost achieved 96% accuracy, demonstrating ECMT's capabilities against complex IoT threats. The integrated approach provides automated, adaptable detection scalable to large IoT deployments through efficient linear models and robust ensemble learning. ECMT addresses concept drift and interpretability via retraining and explanation techniques. Results highlight advanced memory analysis and optimized ML classifiers as a promising solution for robust IoT malware detection despite adversaries' evolving tactics. Further research can extend platform support, harden models against attacks, and refine streaming input. ECMT establishes a foundation for IoT security

With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threat... more With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threats have increased. Identifying malicious attacks in IoT requires advanced techniques tailored to this ecosystem. Existing algorithms have limited effectiveness in detecting obfuscated IoT malware. This study proposes the Elucidating Cybersecurity-promulgated Malware Taxonomy (ECMT) framework, combining memory analysis and ensemble machine learning, to enhance IoT malware categorization. ECMT integrates Support Vector Classification, Quadratic Discriminant Analysis, and AdaBoost on forensic artifacts from memory dumps to improve detection across families like ransomware, spyware, and trojans. ECMT can enable intrusion prevention, information protection, and cybercrime deterrence in IoT environments. Experiments on a balanced dataset indicate AdaBoost achieved 96% accuracy, demonstrating ECMT's capabilities against complex IoT threats. The integrated approach provides automated, adaptable detection scalable to large IoT deployments through efficient linear models and robust ensemble learning. ECMT addresses concept drift and interpretability via retraining and explanation techniques. Results highlight advanced memory analysis and optimized machine learning classifiers as a promising solution for robust IoT malware detection despite adversaries' evolving tactics. Further research can extend platform support, harden models against attacks, and refine streaming input. ECMT establishes a foundation for IoT security by unifying memory forensics, optimized neural architectures, and tailored ensemble learning.

Pulmonary Fibrosis (PF) is an immedicable respiratory condition distinguished by permanent fibrot... more Pulmonary Fibrosis (PF) is an immedicable respiratory condition distinguished by permanent fibrotic alterations in the pulmonary tissue for which there is no cure. Hence, it is crucial to diagnose PF swiftly and precisely. The existing research on deep learning-based pulmonary fibrosis detection methods has limitations, including dataset sample sizes and a lack of standardization in data preprocessing and evaluation metrics. This study presents a comparative analysis of four vision transformers regarding their e_cacy in accurately detecting and classifying patients with Pulmonary Fibrosis and their ability to localize abnormalities within Images obtained from Computerized Tomography (CT) scans. The dataset consisted of 13,486 samples selected out of 24647 from the Pulmonary Fibrosis dataset, which included both PF-positive CT and normal images that underwent preprocessing. The preprocessed images were divided into three sets: the training set, which accounted for 80% of the total pictures; the validation set, which comprised 10%; and the test set, which also consisted of 10%. The vision transformer models, including ViT, MobileViT2, ViTMSN, and BEiT were subjected to training and validation procedures, during which hyperparameters like the learning rate and batch size were fine-tuned. The overall performance of the optimized architectures has been assessed using various performancemetrics to showcase the consistent performance of the fine-tuned model. Regarding performance, ViT has shown superior performance in validation and testing accuracy and loss minimization, specifically for CT images when trained at a single epoch with a tuned learning rate of 0.0001. The results were as follows: validation accuracy of 99.85%, testing accuracy of 100%, training loss of 0.0075, and validation loss of 0.0047. The experimental evaluation of the independently collected data gives empirical evidence that the optimized Vision Transformer (ViT) architecture exhibited superior performance compared to all other optimized architectures. It achieved a flawless score of 1.0 in various standard performance metrics, including Sensitivity, Specificity, Accuracy, F1-score, Precision, Recall, Mathew Correlation Coe_cient (MCC), Precision-Recall Area under the Curve (AUC PR), Receiver Operating Characteristic and Area Under the Curve (ROC-AUC). Therefore, the optimized Vision Transformer (ViT) functions as a reliable diagnostic tool for the automated categorization of individuals with pulmonary fibrosis (PF) using chest computed tomography (CT) scans.

IoT based smart interaction framework for elearning

Proceedings of the 33rd Annual ACM Symposium on Applied Computing, 2018

International Journal of Distributed Sensor Networks, 2019

With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threat... more With the proliferation of connected devices in the Internet of Things (IoT), cybersecurity threats have increased. Identifying malicious attacks in IoT requires advanced techniques tailored to this ecosystem. Existing algorithms have limited effectiveness in detecting obfuscated IoT malware. This study proposes the Elucidating Cybersecurity-promulgated Malware Taxonomy (ECMT) framework, combining memory analysis and ensemble machine learning, to enhance IoT malware categorization. ECMT integrates Support Vector Classification, Quadratic Discriminant Analysis, and AdaBoost on forensic artifacts from memory dumps to improve detection across families like ransomware, spyware, and trojans. ECMT can enable intrusion prevention, information protection, and cybercrime deterrence in IoT environments. Experiments on a balanced dataset indicate AdaBoost achieved 96% accuracy, demonstrating ECMT's capabilities against complex IoT threats. The integrated approach provides automated, adaptable detection scalable to large IoT deployments through efficient linear models and robust ensemble learning. ECMT addresses concept drift and interpretability via retraining and explanation techniques. Results highlight advanced memory analysis and optimized machine learning classifiers as a promising solution for robust IoT malware detection despite adversaries' evolving tactics. Further research can extend platform support, harden models against attacks, and refine streaming input. ECMT establishes a foundation for IoT security by unifying memory forensics, optimized neural architectures, and tailored ensemble learning.