Hammad Afzal

A hybrid routing protocol for wireless sensor networks with mobile sinks

Low power and lossy networks have been an active area of research due to their large number of potential applications in different environments like health, environment monitoring and entertainment domain. Numbers of protocols have been proposed for routing in these networks using metrics like hop count, delay, bandwidth etc. The working group of IETF has done one major contribution in form of a proactive gradient based routing protocol for low power and lossy networks (RPL). However, for a network having few mobile sinks calculating gradients using proactive approach is costly in terms of energy. This paper proposes a hybrid routing protocol for wireless sensor networks with mobile sinks. It proposes a combination of reactive and proactive approach to enhance RPL for efficiently handling movement of multiple sinks. DAGs are only maintained by nodes close to the sink within a certain zone. While the nodes outside the zone use on demand sink discovery to find the closest possible sink, without maintaining DAG. The frequency of zone creation messages and zone sizes can increase or decrease depending on the speed of sink. This helps to decrease the number of retransmissions resulting in low standing cost for maintain DAGs and enhances the network life time especially under average or high mobility of sink.

Improving handwriting based gender classification using ensemble classifiers

A system to predict gender from images of handwriting using textural descriptors.Multiple classifiers to discriminate male and female writings.Classifiers combined using bagging, voting and stacking techniques.Generic and script-independent approach applied to English and Arabic handwritings.Improved results on the QUWI database once compared to state-of-the-art methods. This paper presents a system to predict gender of individuals from offline handwriting samples. The technique relies on extracting a set of textural features from handwriting samples of male and female writers and training multiple classifiers to learn to discriminate between the two gender classes. The features include local binary patterns (LBP), histogram of oriented gradients (HOG), statistics computed from gray-level co-occurrence matrices (GLCM) and features extracted through segmentation-based fractal texture analysis (SFTA). For classification, we employ artificial neural networks (ANN), support vector machine (SVM), nearest neighbor classifier (NN), decision trees (DT) and random forests (RF). Classifiers are then combined using bagging, voting and stacking techniques to enhance the overall system performance. The realized classification rates are significantly better than those of the state-of-the-art systems on this problem validating the ideas put forward in this study.

Towards Creation of Linguistic Resources for Bilingual Sentiment Analysis of Twitter Data

This paper presents an approach towards bi-lingual sentiment analysis of tweets. Social networks being most advanced and popular communication medium can help in designing better government and business strategies. There are a number of studies reported that use data from social networks; however, most of them are based on English language. In this research, we have focused on sentiment analysis of bilingual dataset (English and Roman-Urdu) on topic of national interest (General Elections). Our experiments produced encouraging results with 76% of tweet’s sentiment strength classified correctly. We have also created a bi-lingual lexicon that stores the sentiment strength of English and Roman Urdu terms. Our lexicon is available at: https://sites.google. com/a/mcs.edu.pk/codteem/biling_senti

Mining semantic networks of bioinformatics e-resources from the literature

BackgroundThere have been a number of recent efforts (e.g. BioCatalogue, BioMoby) to systematically catalogue bioinformatics tools, services and datasets. These efforts rely on manual curation, making it difficult to cope with the huge influx of various electronic resources that have been provided by the bioinformatics community. We present a text mining approach that utilises the literature to automatically extract descriptions and semantically profile bioinformatics resources to make them available for resource discovery and exploration through semantic networks that contain related resources.ResultsThe method identifies the mentions of resources in the literature and assigns a set of co-occurring terminological entities (descriptors) to represent them. We have processed 2,691 full-text bioinformatics articles and extracted profiles of 12,452 resources containing associated descriptors with binary and tf*idf weights. Since such representations are typically sparse (on average 13.77 features per resource), we used lexical kernel metrics to identify semantically related resources via descriptor smoothing. Resources are then clustered or linked into semantic networks, providing the users (bioinformaticians, curators and service/tool crawlers) with a possibility to explore algorithms, tools, services and datasets based on their relatedness. Manual exploration of links between a set of 18 well-known bioinformatics resources suggests that the method was able to identify and group semantically related entities.ConclusionsThe results have shown that the method can reconstruct interesting functional links between resources (e.g. linking data types and algorithms), in particular when tf*idf-like weights are used for profiling. This demonstrates the potential of combining literature mining and simple lexical kernel methods to model relatedness between resource descriptors in particular when there are few features, thus potentially improving the resource description, discovery and exploration process. The resource profiles are available at http://gnode1.mib.man.ac.uk/bioinf/semnets.html

Mining Semantic Descriptions of Bioinformatics Web Resources from the Literature

A number of projects (myGrid, BioMOBY, etc.) have recently been initiated in order to organise emerging bioinformatics Web Services and provide their semantic descriptions. They typically rely on manual curation efforts. In this paper we focus on a semi-automated approach to mine semantic descriptions from the bioinformatics literature. The method combines terminological processing and dependency parsing of journal articles, and applies information extraction techniques to profile Web services using informative textual passages, related ontological annotations and service descriptors. Service descriptors are terminological phrases reflecting related concepts (e.g. tasks, approaches, data) and/or specific roles (e.g. input/output parameters, etc.) of the associated resource classes (e.g. algorithms, databases, etc.). They can be used to facilitate subsequent manual description of services, but also for providing a semantic synopsis of a service that can be used to locate related services. We present a case-study involving full text articles from the BMC Bioinformatics journal. We illustrate the potential of natural language processing not only for mining descriptions of known services, but also for discovering new services that have been described in the literature.

An Approach for Secure Semantic Data Integration at Data as a Service (DaaS) Layer

 Abstract—Data integration provides a uniform view of a set of heterogeneous data sources and facilitates users to query without any knowledge of the underlying heterogeneous data sources. In current era, Service Oriented Architecture and Cloud computing together has enabled users to access services over the Internet at a low cost. Cloud computing model provides a layer which is responsible for providing data to the other layers and services i.e., Data as a service (DaaS) layer. The issue of providing an integrated view of data can be handled using Semantic data; the data stored in a way that is understandable by machines and integratable without human intervention. However, integrating data using semantic web technology without enforcing any access management will raise privacy and confidentiality concerns. Different data owners store data in heterogeneous format based on their requirements. This leads to the data interoperability problem. This research proposes a framework that would allow data from different sources to be integrated using the concept of semantic data, thus resolving the issue of interoperability and also devises an access control system for defining explicit privacy constraints.

Vocabulary of Quranic Concepts: A semi-automatically created terminology of Holy Quran

The identification and organization of terminology is the foremost step while organizing the domain knowledge for any domain as it is the terms and their inter-relationships that define the conceptual knowledge base. Quran, comprising the divine words of wisdom has been considered and used as prime source of knowledge and guidance for Muslims throughout the world for fourteen centuries. The concepts/topics discussed in Quran have been organized/indexed by many scholars which are used by Muslims who use them to search for guidance regarding various issues of daily life. In current era of information technology, various search services for Quranic topics are available online. They mostly use the terminologies (concepts hierarchy) manually built by scholars. In our work, we have used a semi-automatic approach to identify important concepts/topics from six English translations of Quran, and organized them into a hierarchical structure, named as Vocabulary of Quranic Concepts (VQC). CNC Value method of term recognition is used to identify significant concepts, which are then manually analyzed by domain expert, and are then organized into a hierarchy using the term-head principle. Due to extreme sensitivity of this work, complete automation of system is avoided and outcomes at all steps are manually analyzed. Currently, we have developed a vocabulary from translation of only second chapter of Quran (Al-Bakara). VQC is available at: https://sites.google.com/a/mcs.edu.pk/codteem/projects/qwn

Creation of Bi-lingual Social Network Dataset Using Classifiers

This paper presents an approach towards creation of topic focused short text (social data) dataset using classification. With emerging use of internet, social networks have turned as the most advanced tool for information sharing among communities. Different communities from different backgrounds use globally renowned social networks often using and promoting their own cultures and languages. Hence, such information exchange turns social networks into multi-lingual information hubs. There are a number of behavioral and demographic oriented analytical studies reported that use data from social networks, but most of the studies are performed using English. In this study, we have focused on development of topic oriented bi-lingual dataset that can be used as corpus to perform further analytical studies. The languages focused are English and Roman-Urdu (which is spoken by about 8 million active users of social network). The main contribution is bi-lingual classifier which is used to create English and Roman-Urdu classified tweets dataset.

Spam filtering of bi-lingual tweets using machine learning

During recent years, usage of social media has increased enormously. Billions of users use Twitter, Youtube etc which has resulted in the increase in spams as well. Spammers use spam accounts and target users on online social media. Whether a user accesses this social media through smart-phone or web, he/she is prone to the spammers on social media websites. This paper analyses different classification techniques that are currently being used in spam filtering in the context of social media. The contents of tweets are unique in nature, and are different from emails due to their less content so some techniques used in emails might be effective while some might not be effective. Moreover, the conversations on social media often comprises of short-forms/slangs and incorrect spellings. Usage of social media has also become popular in local/regional languages. One such language is Urdu which is common in subcontinent Indo-Pak and is written using English alphabets. We have performed spam classification for Roman Urdu tweets, collected from five major cities of Pakistan. Some of the most commonly used algorithms and techniques for spam classification are discussed and evaluated on English and Roman Urdu tweets from Pakistan in this paper.

Contributions to the study of bi-lingual Roman Urdu SMS spam filtering

With the increased usage of internet and mobile phones, number of spams has also increased in both these areas. The Spam in both these areas is an increasing threat and sometimes cause huge financial as well as data/confidentiality loss. Therefore, actions need to be taken to stop these spams on both media. This paper analyses various techniques that are currently being used in Spam filtering in the context of mobile text messages. The contents of SMS are unique in nature so some techniques might be effective while some might not be. Some of mostly used algorithms and techniques are discussed in this paper. Furthermore, we have performed automatic spam filtering using machine learning algorithms on Roman Urdu text messages and achieved an accuracy of 92.2% on a manually curated corpus of 8449 messages. The SMS corpus has also been made available for future research works.