Teng-Sheng Moh

On data gathering protocols for in-body biomedical sensor networks

This paper investigates the effectiveness of data gathering protocols for in-body biomedical sensor networks. We studied the performance of representatives from each of three major protocol categories: (1) low energy adaptive clustering hierarchy (LEACH), a cluster-based protocol, (2) power efficient gathering for sensory information systems (PEGASIS), a chain-based protocol, and (3) hybrid indirect transmissions (HIT), a hybrid of chains and clusters. First, the ability of each protocol to perform in-network source separation was judged. We consider a human-machine interaction application in which implanted bio-sensors communicate motor unit actions of human muscles to a remote computer. Motor unit action potentials (MUAPs) were modeled, and source separation and recovery at each sensor was simulated. We compare the performance of HIT and LEACH in terms of signal distortion ratios and the energy costs of fusion and communication. Second, we report on the efficiency of each protocol for in-body data collection, using Gupta et als propagation loss model for biomedical applications (PMBA) - an accurate model of power loss due to signal absorption by the human body. We investigate the effectiveness of HIT, LEACH, and PEGASIS under this model, and compare their performance in terms of energy efficiency and network lifetime

Can You Judge a Man by His Friends? - Enhancing Spammer Detection on the Twitter Microblogging Platform Using Friends and Followers

As online social networks acquire a larger user base, they also become more interesting targets for spammers. Spam can take very different forms on social web sites and can not always be detected by analyzing textual content. However, the platform’s social nature also offers new ways of approaching the spam problem. In this work we analyze a user’s friends and followers to gain information on him. Next, we evaluate them using different metrics to determine the amount of trust his peers give him. We use the Twitter microblogging platform for this case study.

Privacy and security in biomedical applications of wireless sensor networks

Wireless sensor network applications in healthcare and biomedical technology have received increasing attention, while associated security and privacy issues remain open areas of consideration. The relevance of this technology to our growing elderly population, as well as our increasingly over-crowded and attention-drained healthcare systems, is promising. However, prior to the emergence of these systems as a ubiquitous technology, healthcare providers and regulatory agencies must determine an acceptable level of security and privacy. This paper will review biomedical applications of wireless sensor networks, identify security and privacy issues to be addressed, and note some of the proposed methods for securing these systems.

Detecting Web Attacks Using Multi-stage Log Analysis

Web-based applications have gained universal acceptance in every sector of lives, including social, commercial, government, and academic communities. Even with the recent emergence of cloud technology, most of cloud applications are accessed and controlled through web interfaces. Web security has therefore continued to be fundamentally important and extremely challenging. One major security issue of web applications is SQL-injection attacks. Most existing solutions for detecting these attacks use log analysis, and employ either pattern matching or machine learning methods. Pattern matching methods can be effective, dynamic, they however cannot detect new kinds of attacks. Supervised machine learning methods can detect new attacks, yet they need to rely on an off-line training phase. This work proposes a multi-stage log analysis architecture, which combines both pattern matching and supervised machine learning methods. It uses logs generated by the application during attacks to effectively detect attacks and to help preventing future attacks. The architecture is described in detail, a proof-of-concept prototype is implemented and hosted on Amazon AWS, using Kibana for pattern matching and Bayes Net for machine learning. It is evaluated on 10,000 logs for detecting SQL injection attacks. Experiment results show that the two-stage system has combined the advantages of both systems, and has substantially improved the detection accuracy. The proposed work is significant in advancing web securities, while the multi-stage log analysis concept would be highly applicable to many intrusion detection applications.

DBSCAN on Resilient Distributed Datasets

DBSCAN is a well-known density-based data clustering algorithm that is widely used due to its ability to find arbitrarily shaped clusters in noisy data. However, DBSCAN is hard to scale which limits its utility when working with large data sets. Resilient Distributed Datasets (RDDs), on the other hand, are a fast data-processing abstraction created explicitly for in-memory computation of large data sets. This paper presents a new algorithm based on DBSCAN using the Resilient Distributed Datasets approach: RDD-DBSCAN. RDD-DBSCAN overcomes the scalability limitations of the traditional DBSCAN algorithm by operating in a fully distributed fashion. The paper also evaluates an implementation of RDD-DBSCAN using Apache Spark, the official RDD implementation.

Evaluation of dynamic tree-based data gathering algorithms for wireless sensor networks

One major challenge in sensor networks is to maximize network life under the constraint of extremely limited power supply. Thus, two important design issues of routing and data-gathering protocols are 1) minimizing energy consumption in sensor nodes and 2) adapting to node failures. This paper studies two tree-based data gathering protocols, based on distributed versions of shortest path tree (SPT) and maximum leaf tree (MLT) algorithms. Furthermore, the two distributed algorithms are extended to be dynamic and robust. A localized tree-reconstruction scheme, localized flooding algorithm, is added to handle joining and leaving (death) of sensor nodes. Accurate energy consumption has been modeled for both leaf-nodes and intermediate nodes, when sending and receiving control and data packets. The resulting dynamic algorithms are fast to adapt network changes. Performance is evaluated through detailed simulation. Comparing with MLT, due to its simplicity and smaller number of control message exchanges, SPT achieves better energy efficiency and less delay in tree constructions, data transmissions, and dynamic tree reconstructions

A running time improvement for the two thresholds two divisors algorithm

Chunking algorithms play an important role in hash-based data de-duplication systems. The Basic Sliding Window (BSW) algorithm is the first prototype of a content-based chunking algorithm that can handle most types of data. The Two Thresholds Two Divisors (TTTD) algorithm was proposed to improve the BSW algorithm by controlling the chunk-size variations. We conducted a series of systematic experiments to evaluate the performances of these two algorithms. We also proposed a new improvement for the TTTD algorithm. Our new approach reduced about 6% of the running time and 50% of the large-sized chunks, and also brought other significant benefits.

Path-Based Reputation System for MANET Routing

Most existing reputation systems in mobile ad hoc networks (MANET) consider only node reputations when selecting routes. Therefore, reputation and trust are only ensured within a one-hop distance when routing decisions are made. This often fails to provide the most reliable, trusted route. In this paper we propose a system that is based on path reputation , which is computed from the reputation and trust values of each and every node in the route. This greatly enhances the reliability of the resulting routes. The system is simulated on top of the AODV (Ad-hoc On-demand Distance Vector) routing protocol. It is effective at detecting misbehaving nodes, including selfishness and worm-hole attacks. It greatly improves network throughput in the midst of malicious nodes and requires very limited message overhead. To our knowledge, this is the first path-based reputation system proposal that is applicable to non-source-based routing schemes.

A framework for fast-feedback opinion mining on Twitter data streams

This paper focuses on the computational infrastructure for fast-feedback opinion mining. This calls for a versatile platform to handle all the possible problems arisen from mining data streams of a social networking site. In particular, we consider the difficulty of getting customer feedbacks faced by companies that produce free software. This is especially challenging since, when encountering buggy software, customers would just switch to another free software with similar functionality without providing any feedback. Our framework makes use of real-time Twitter data stream. These data streams are filtered and analyzed and fast feedback is obtained through opinion mining. The framework is built upon Apache Hadoop to deal with huge volume of data streamed from Twitter. The experiments have shown an 84% accuracy in the sentimental analysis. Our framework is therefore able to provide fast, valuable feedbacks to companies.

On Multi-tier Sentiment Analysis Using Supervised Machine Learning

Document management and Information Retrieval tasks have rapidly increased due to the availability of digital documents anytime, any place. The need for automatic extraction of document information has become prominent in information organization and knowledge discovery. Text Classification is one such solution, where in the natural language text is assigned to one or more predefined categories based on the content. This work focuses on sentiment analysis, also known as opinion mining. It is a way of automatically extracting and analyzing the emotions and opinions, and not facts, of messages and posts. A multi-tier classification architecture is proposed, which consists of major modules such as data cleaning and pre-processing, feature selection, and classifier training that includes a multi-tier prediction model. The architecture and its components are carefully described. Four classifiers (Naïve Bayes, SVM, Random Forest, and SGD) are used in the experiments, which evaluate the performance of the proposed multi-tier architecture by analyzing the sentiments and opinions of 150,000 movie reviews. Results have shown that the multi-tier model is able to significantly improve prediction accuracy over the single-tier model by more than 10%, the improvement is significant when customized dictionary is used. We believe that the proposed multi-tier classification architecture, with the various feature selection techniques described and used, are significant, and are readily applicable to many other areas of sentiment analysis.