Yousra Asim

Community Detection in Networks using Node Attributes and Modularity

Community detection in network is of vital importance to find cohesive subgroups. Node attributes can improve the accuracy of community detection when combined with link information in a graph. Community detection using node attributes has not been investigated in detail. To explore the aforementioned idea, we have adopted an approach by modifying the Louvain algorithm. We have proposed Louvain-AND-Attribute (LAA) and Louvain-OR-Attribute (LOA) methods to analyze the effect of using node attributes with modularity. We compared this approach with existing community detection approaches using different datasets. We found the performance of both algorithms better than Newman’s Eigenvector method in achieving modularity and relatively good results of gain in modularity in LAA than LOA. We used density, internal and external edge density for the evaluation of quality of detected communities. LOA provided highly dense partitions in the network as compared to Louvain and Eigenvector algorithms and close values to Clauset. Moreover, LOA achieved few numbers of edges between communities.

Software-Defined Networks (SDNs) and Internet of Things (IoTs): A Qualitative Prediction for 2020

The Internet of Things (IoT) is imminent technology grabbing industries and research attention with a fast stride. Currently, more than 15 billion devices are connected to the Internet and this number is expected to reach up to 50 billion by 2020. The data generated by these IoT devices are immensely high, creating resource allocation, flow management and security jeopardises in the IoT network. Programmability and centralised control are considered an alternative solution to address IoT issues. On the other hand, a Software Define Network (SDN) provides a centralised and programmable control and management for the underlying network without changing existing network architecture. This paper surveys the state of the art on the IoT integration with the SDN. A comprehensive review and the generalised solutions over the period 2010-2016 is presented for the different communication domains. Furthermore, a critical review of the IoT and the SDN technologies, current trends in research and the futuristic contributing factors form part of the paper. The comparative analysis of the existing solutions of SDN based IoT implementation provides an easy and concise view of the emerging trends. Lastly, the paper predicts the future and presents a qualitative view of the world in 2020.

A Comparison of Collaborative Access Control Models

Collaborative environments need access control to data and resources to increase working cooperation efficiently yet effectively. Several approaches are proposed and multiple access control models are recommended in this domain. In this paper, four Role-Based Access Control (RBAC) based collaborative models are selected for analysis and comparison. The standard RBAC model, Team-based Access Control (TMAC) model, Privacy-aware Role-Based Access Control (P-RBAC) model and Dynamic Sharing and Privacy-aware RBAC (DySP-RBAC) model are used for experiments. A prototype is developed for each of these models and pros and cons of these models are discussed. Performance and sharing parameters are used to compare these collaborative models. The standard RBAC model is found better by having a quick response time for queries as compared to other RBAC models. The DySP-RBAC model outperforms other models by providing enhanced sharing capabilities.

Constraints in the IoT: The World in 2020 and Beyond

The Internet of Things (IoT), often referred as the future Internet; is a collection of interconnected devices integrated into the world-wide network that covers almost everything and could be available anywhere. IoT is an emerging technology and aims to play an important role in saving money, conserving energy, eliminating gap and better monitoring for intensive management on a routine basis. On the other hand, it is also facing certain design constraints such as technical challenges, social challenges, compromising privacy and performance tradeoffs. This paper surveys major technical limitations that are hindering the successful deployment of the IoT such as standardization, interoperability, networking issues, addressing and sensing issues, power and storage restrictions, privacy and security, etc. This paper categorizes the existing research on the technical constraints that have been published in the recent years. With this categorization, we aim to provide an easy and concise view of the technical aspects of the IoT. Furthermore, we forecast the changes influenced by the IoT. This paper predicts the future and provides an estimation of the world in year 2020 and beyond.

Personal vs. know-how contacts: which matter more in wiki elections?

The use of online social media is also connected with the real world. A very common example of this is the effect of social media coverage on the chances of success of elections. Previous literature has identified that the outcome of elections can often be predicted based on online public discussions. These discussions can be across various online social network with a special focus on the candidates own accounts. Among many other forms of social media, Wikipedia is a very widely-used self-organizing information resource. The management and administration of Wikipedia is performed using special users which are elected by means of online public elections. In other words, the results of these elections pose as an emergent outcome of a large-scale self-organized opinion formation process. However, due to dynamical, and non-linear interactions besides the presence of mutual dependencies between election participants, a statistical analysis of this data can both be cumbersome as well as inefficient in terms of information extraction. We believe that social network analysis is a more appropriate alternative. It allows for the identification of local and global patterns, identification of influential nodes as well as the contacts involved in the influence. In general, this particular analytic technique can help in examining the internal complex network dynamics. In the current paper, we investigates whether personal contacts matter more than know-how contacts in wiki election nominations and voting participation. We employ the use of standard social network analysis tools such as Pajek and Gephi. The presented work demonstrates the significance of personal contacts over know-how contacts of a person in online elections. We have discovered that personal contacts, i.e. immediate neighbors (based on degree centrality) and neighborhood (k-neighbors) of a person have a positive effect on a person’s nomination as an administrator and also contribute to the active participation of voters in voting. Moreover, know-how contacts, analyzed by means of measures such as betweenness and closeness centralities, have a relatively insignificant effect on the selection of a person. However, know-how contacts, measured in terms of betweenness centrality can positively contribute only to the voting process—primarily due to the role played in passing information around the network. These contacts, also measured in terms of influence domain and PageRank, can play a vital role in the selection of an admin. Additionally, such contacts have a positive association with the voting process in terms of reachability and brokerage roles.

A multi-modal, multi-atlas-based approach for Alzheimer detection via machine learning

The use of biomarkers for early detection of Alzheimers disease (AD) improves the accuracy of imaging‐based prediction of AD and its prodromal stage that is mild cognitive impairment (MCI). Brain parcellation‐based computer‐aided methods for detecting AD and MCI segregate the brain in different anatomical regions and use their features to predict AD and MCI. Brain parcellation generally is carried out based on existing anatomical atlas templates, which vary in the boundaries and number of anatomical regions. This works considers dividing the brain based on different atlases and combining the features extracted from these anatomical parcellations for a more holistic and robust representation. We collected data from the ADNI database and divided brains based on two well‐known atlases: LONI Probabilistic Brain Atlas (LPBA40) and Automated Anatomical Labeling (AAL). We used baselines images of structural magnetic resonance imaging (MRI) and 18F‐fluorodeoxyglucose positron emission tomography (FDG‐PET) to calculate average gray‐matter density and average relative cerebral metabolic rate for glucose in each region. Later, we classified AD, MCI and cognitively normal (CN) subjects using the individual features extracted from each atlas template and the combined features of both atlases. We reduced the dimensionality of individual and combined features using principal component analysis, and used support vector machines for classification. We also ranked features mostly involved in classification to determine the importance of brain regions for accurately classifying the subjects. Results demonstrated that features calculated from multiple atlases lead to improved performance compared to those extracted from one atlas only.

Rule Adaptation in Collaborative Working Environments using RBAC Model

Collaborative Working Environments (CWEs) are getting prominence these days. With the increase in the use of collaboration tools and technologies, a lot of sharing and privacy issues have also emerged. Due to its dynamic nature, a CWE needs to adapt the changes into accordingly. In this paper, we have implemented the Adaptive Dynamic Sharing and Privacy-aware Role Based Access Control (Adaptive DySP-RBAC) model which provides user’s information privacy to dynamically adapt the changes occurring in the system at any time. The proposed model has been implemented as a prototype and tested. Results have shown that our system efficiently and effectively adapts access rules according to the changes happening in a CWE along with preserving the user’s information privacy in the system.

Significance of machine learning algorithms in professional blogger's classification

Abstract Outreach of internet has opened new horizons for the people who want quick and widespread dissemination of their ideas, and the tool to do so is blogging. Bloggers can broadly be classified into two groups: professional and non-professional bloggers. As for professional bloggers, there are many factors that influence individuals to opt this profession. This study, with the help of an online dataset, attempts to identify such factors. Data analysis was made by using decision tree algorithms, lazy learning algorithms and ensembling methods. Nearest-neighbour classifier (IB1) and RandomForest have results with 85% accuracy and 84.8% precision for classification. The proof of concept is provided for result validation. The causes behind the varying performance of algorithms are elaborated. The factors that influence a blogger to behave professionally are identified based on the classifier with the best results.

A community-centric access control scheme for Online Social Networks

Online Social Network facilitate its users to form groups for sharing their personal information, interests and activities with their friends, relatives and others. Such type of information disclosure can cause problems for people in terms of security. Access of different group members is controlled in Online Social Networks to minimize such security risks to personal information and user resources. Researchers have used social network actor-relationship graph to control the access based on relationship type, trust and attributes. This paper discusses available techniques in CBAC, ABAC and ReBAC to provide security to social network data and objects. A community-centric access control model using brokerage roles, attributes, and trust techniques is presented with its formal description.

Automated Colon Cancer Detection Using Structural and Morphological Features

Traditionally, colon cancer is diagnosed using microscopic analysis of histopathological colon samples. The manual examination of tissue specimens is not only time consuming, but is also subjective, and depends upon a few factors such as experience and work-load of the histopathologist. Therefore, research community is constantly putting efforts in developing automated colon cancer diagnostic systems, which can provide reliable second opinion to the histopathologists. Colon biopsy image based classification is one of such computer-aided diagnostic technique that can help in quantifying the differences in the structure of normal and malignant colon tissues without needing the subjective involvement of histopathologists. In this work, we propose a computer-aided diagnostic technique that models the differences in the regular organization of normal colon tissues and irregular structure of malignant colon tissues in terms of a few features such as least square distances, elliptic Fourier descriptors (EFDs) and morphological features. These features are extracted from each colon biopsy image, and are given as input to classifier. An ensemble of SVM kernels, based on majority voting, has been developed for classification of samples into normal and malignant classes. Features are also combined to develop an information rich hybrid feature vector, which is also used for the classification. The proposed method has been tested on a colon biopsy image based dataset, and performance has been observed in terms of accuracy, sensitivity, specificity, receiver operating characteristics (ROC) curves, area under the curve (AUC), and Kappa statistics. It has been observed that each feature type performs reasonably well. Further, ensemble classifier and the hybrid feature vector have shown better performance compared to individual features and the individual classifiers, respectively.