Computer Sciences Commons^™

An Open Source Advertisement Server, Pushkar Umaranikar

Master's Projects

This report describes a new online advertisement system and its implementation for the Yioop open source search engine. This system was implemented for my CS298 project. It supports both selling advertisements and displaying them within search results. The selling of advertisement is done using a novel auction system, which we describe in this paper. With this auction system, it is possible to create an advertisement, attach keywords to it, and add it to the advertisement inventory. An advertisement is displayed on a search results page if the search keyword matches the keywords attached to the advertisement. Display of advertisements is …

Context-Based Autosuggest On Graph Data, Hai Nguyen

Master's Projects

Autosuggest is an important feature in any search applications. Currently, most applications only suggest a single term based on how frequent that term appears in the indexed documents or how often it is searched upon. These approaches might not provide the most relevant suggestions because users often enter a series of related query terms to answer a question they have in mind. In this project, we implemented the Smart Solr Suggester plugin using a context-based approach that takes into account the relationships among search keywords. In particular, we used the keywords that the user has chosen so far in the …

A Scalable Search Engine Aggregator, Pooja Mishra

Master's Projects

The ability to display different media sources in an appropriate way is an integral part of search engines such as Google, Yahoo, and Bing, as well as social networking sites like Facebook, etc. This project explores and implements various media-updating features of the open source search engine Yioop [1]. These include news aggregation, video conversion and email distribution. An older, preexisting news update feature of Yioop was modified and scaled so that it can work on many machines. We redesigned and modified the user interface associated with a distributed news updater feature in Yioop. This project also introduced a video …

Driver Telematics Analysis, Karthik Vakati

Master's Projects

For automobile insurance firms, telemetric analysis represents a valuable and growing way to identify the risk associated with each driver. The pricing decisions of an insurer are best accounted for if they are made considering the driver’s behavior instead of just the vehicle characteristics and the best way to understand a driver’s behavior is to leverage the telemetric analysis. Decisions made on such factors can eventually lead to increased premium or reduced liability for unsafe or reckless drivers and can also help in transitioning the burden to the policies that lead to increased liability.

The dataset provided for this project …

Maximizing The Speed Of Influence In Social Networks, Yubo Wang

Master's Projects

Influence maximization in social networks is the problem of selecting a limited

size of influential users as seed nodes so that the influence from these seed nodes can propagate to the largest number of other nodes in the network. Previous studies in influence maximization focused on three areas, i.e., designing propagation models, improving algorithms of seed-node selection and exploiting the structure of social networks. However, most of these studies ignored the time constraint in influence propagation. In this paper, I studied how to maximize influence propagation in a given time, i.e., maximizing the speed of influence propagation in social networks. …

Using Neural Networks For Image Classification, Tim Kang

Master's Projects

This paper will focus on applying neural network machine learning methods to images for the purpose of automatic detection and classification. The main advantage of using neural network methods in this project is its adeptness at fitting non­linear data and its ability to work as an unsupervised algorithm. The algorithms will be run on common, publically available datasets, namely the MNIST and CIFAR­10, so that our results will be easily reproducible.

Static Analysis Of Malicious Java Applets, Nikitha Ganesh

Master's Projects

In this research, we consider the problem of detecting malicious Java applets, based on static analysis. In general, dynamic analysis is more informative, but static analysis is more efficient, and hence more practical. Consequently, static analysis is preferred, provided we can obtain results comparable to those obtained using dynamic analysis. We conducted experiments with the machine learning technique, Hidden Markov Model (HMM). We show that in some cases a static technique can detect malicious Java applets with greater accuracy than previously published research that relied on dynamic analysis.

Sociobot: Twitter For Command And Control Of A Botnet, Ismeet Kaur Makkar

Master's Projects

A botnet is a collection of computers controlled by a botmaster, often used for malicious activity. Social network provides an ideal medium for botnets to spread their reach. In this research, we develop and analyze a botnet that uses Twitter for its command and control channel. We use this botnet to perform a distributed denial of service attack on a web server, and we utilize the biological epidemic models to analyze the spread of the botnet using Twitter.

Combining Dynamic And Static Analysis For Malware Detection, Anusha Damodaran

Master's Projects

Well-designed malware can evade static detection techniques, such as signature scanning. Dynamic analysis strips away one layer of obfuscation and hence such an approach can potentially provide more accurate detection results. However, dynamic analysis is generally more costly than static analysis. In this research, we analyze the effectiveness of using dynamic analysis to enhance the training phase, while using only static techniques in the detection phase. Relative to a fully static approach, the additional overhead is minimal, since training is essentially one-time work.

Using Probabilistic Graphical Models To Solve Np-Complete Puzzle Problems, Fengjiao Wu

Master's Projects

Probabilistic Graphical Models (PGMs) are commonly used in machine learning to solve problems stemming from medicine, meteorology, speech recognition, image processing, intelligent tutoring, gambling, games, and biology. PGMs are applicable for both directed graph and undirected graph. In this work, I focus on the undirected graphical model. The objective of this work is to study how PGMs can be applied to find solutions to two puzzle problems, sudoku and jigsaw puzzles. First, both puzzle problems are represented as undirected graphs, and then I map the relations of nodes to PGMs and Belief Propagation (BP). This work represents the puzzle grid …

Comparative Analysis Of Particle Swarm Optimization Algorithms For Text Feature Selection, Shuang Wu

Master's Projects

With the rapid growth of Internet, more and more natural language text documents are available in electronic format, making automated text categorization a must in most fields. Due to the high dimensionality of text categorization tasks, feature selection is needed before executing document classification. There are basically two kinds of feature selection approaches: the filter approach and the wrapper approach. For the wrapper approach, a search algorithm for feature subsets and an evaluation algorithm for assessing the fitness of the selected feature subset are required. In this work, I focus on the comparison between two wrapper approaches. These two approaches …

Operational Semantics For Featherweight Lua, Hanshu Lin

Master's Projects

Lua is a small, embedded language to provide scripting in other languages. De- spite a clean, minimal syntax, it is still too complex for formal reasoning because of some syntactic sugar or specific syntax structures in Lua.

This thesis develops Featherweight Lua (FWLua), following the tradition of lan- guages like Featherweight Java[1] and Featherweight JavaScript[2]. The goal is to develop a core of language features that, while remaining simple enough for formal reasoning, also remain faithful to the central characteristics of the language. Specifi- cally for Lua, the core features that are essential for our modeling include:

∙ First-class functions …

Using Hidden Markov Models To Detect Dna Motifs, Santrupti Nerli

Master's Projects

During the process of gene expression in eukaryotes, mRNA splicing is one of the key processes carried out by a complex called spliceosome. Spliceosome guarantees proper removal of introns and joining of exons before the translation process. Precise splicing is essential for the production of functional proteins. Spliceosome detects specific sequence motifs within an mRNA sequence called splice sites. Two of the splice sites are the 5’ and 3’ sites that border all the introns. Normal splicing process if disrupted by mutation may lead to fatal diseases. In this work, we predict splice sites in a human genome using hidden …

Financial Ratio Analysis For Stock Price Movement Prediction Using Hybrid Clustering, Tom Tupe

Master's Projects

We have gathered over 3100 annual financial reports for 500 companies listed on the S&P 500 index, where the main goal was to select and give proper weights to the various pieces of quantitative data to maximize clustering results and improve prediction results over previous work by [Lin et al. 2011]. Various financial ratios, including earnings per share surprise percentages were gathered and analyzed. We proposed and used two types, correlation based ratios and causality based ratios. An extension to the classification scheme used by [Lin et al. 2011] was proposed to more accurately classify financial reports, together with a …

A Hash-Cash Based Music Streaming Payment System, Timothy Chen

Master's Projects

This project develops a hash-cash based, streaming music payment system. In our system, musicians are paid based on how long their works are listened to. Artists can upload their works to our proof-of-concept service so that people can discover and listen to them. While their works are being listened to, a mining process is run in parallel. The mining process discovers a “listening coin” based on the hash-cash algorithm. Users of our service would pay a monthly fee to access the music library. The monthly fees are then distributed to all artists proportionate to the number of virtual coins they …

Masquerade Detection Using Singular Value Decomposition, Sweta Vikram Shah

Master's Projects

Information systems and networks are highly susceptible to attacks in the form of intrusions. One such attack is by the masqueraders who impersonate legitimate users. Masqueraders can be detected in anomaly based intrusion detection by identifying the abnormalities in user behavior. This user behavior is logged in log files of different types. In our research we use the score based technique of Singular Value Decomposition to address the problem of masquerade detection on a unix based system. We have data collected in the form of sequential unix commands ran by 50 users. SVD is a linear algebraic technique, which has …

Spartan Web Application Firewall, Brian C. Lee

Master's Projects

Computer security is an ongoing issue and attacks are growing more sophisit- cated. One category of attack utilizes cross-site scripting (XSS) to extract confiden- tial data such as a user’s login credential’s without the knowledge of either the user nor the web server by utilizing vulnerabilities on web pages and internet browsers. Many people develop their own web applications without learning about or having good coding practices or security in mind. Web application firewalls are able to help but can be enhanced to be more effective than they currently are at detecting re- flected XSS attacks by analyzing the request …

A Content-Sensitive Wiki Help System, Eswara Satya Pavan Rajesh Pinapala

Master's Projects

Context-sensitive help is a software application component that enables users to open help pertaining to their state, location, or the action they are performing within the software. Context-sensitive “wiki” help, on the other hand, is help powered by a wiki system with all the features of context-sensitive help. A context-sensitive wiki help system aims to make the context-sensitive help collaborative; in addition to seeking help, users can directly contribute to the help system. I have implemented a context-sensitive wiki help system into Yioop, an open source search engine and software portal created by Dr. Chris Pollett, in order to measure …

A Smart Web Crawler For A Concept Based Semantic Search Engine, Vinay Kancherla

Master's Projects

The internet is a vast collection of billions of web pages containing terabytes of information arranged in thousands of servers using HTML. The size of this collection itself is a formidable obstacle in retrieving information necessary and relevant. This made search engines an important part of our lives. Search engines strive to retrieve information as relevant as possible to the user. One of the building blocks of search engines is the Web Crawler. A web crawler is a bot that goes around the internet collecting and storing it in a database for further analysis and arrangement of the data.

The …

Optimization Of Scheduling And Dispatching Cars On Demand, Vu Tran

Master's Projects

Taxicab is the most common type of on-demand transportation service in the city because its dispatching system offers better services in terms of shorter wait time. However, the shorter wait time and travel time for multiple passengers and destinations are very considerable. There are recent companies implemented the real-time ridesharing model that expects to reduce the riding cost when passengers are willing to share their rides with the others. This model does not solve the shorter wait time and travel time when there are multiple passengers and destinations. This paper investigates how the ridesharing can be improved by using the …

Parallelized Rigid Body Dynamics, John Calvin Linford

Master's Projects

Physics engines are collections of API-like software designed for video games, movies and scientific simulations. While physics engines often come in many shapes and designs, all engines can benefit from an increase in speed via parallelization. However, despite this need for increased speed, it is uncommon to encounter a parallelized physics engine today. Many engines are long-standing projects and changing them to support parallelization is too costly to consider as a practical matter. Parallelization needs to be considered from the design stages through completion to ensure adequate implementation. In this project we develop a realistic approach to simulate physics in …

Concept Based Semantic Search Engine, Pradeep Roy

Master's Projects

In the current day and age, search engines are the most relied on and critical ways to find out information on the World Wide Web (W3). With the ushering in of Big Data, traditional search engines are becoming inept and inadequate at dishing out relevant pages. It has become increasingly difficult to locate meaningful results from the mind boggling list of returns typical of returned search queries. Keywords, often times, alone cannot capture the intended concept with high precision. These and associated issues with the current search engines call for a more powerful and holistic search engine capability. The current …

Comparative Analysis Of Naive Bayes And Tree Augmented Naive Bayes Models, Harini Padmanaban

Master's Projects

Naive Bayes and Tree Augmented Naive Bayes (TAN) are probabilistic graphical models used
for modeling huge datasets involving lots of uncertainties among its various interdependent
feature sets. Some of the most common applications of these models are image segmentation,
medical diagnosis and various other data clustering and data classification applications. A
classification problem deals with identifying to which category a particular instance belongs to,
based on previous knowledge acquired by analysis of various such instances. The instances are
described using a set of variables called attributes or features. A Naive Bayes model assumes that
all the attributes of an instance …

A Tunable Workflow Scheduling Algorithm Based On Particle Swarm Optimization For Cloud Computing, Kai Wu

Master's Projects

Cloud computing provides a pool of virtualized computing resources and adopts pay-per-use model. Schedulers for cloud computing make decision on how to allocate tasks of workflow to those virtualized computing resources. In this report, I present a flexible particle swarm optimization (PSO) based scheduling algorithm to minimize both total cost and makespan. Experiment is conducted by varying computation of tasks, number of particles and weight values of cost and makespan in fitness function. The results show that the proposed algorithm achieves both low cost and makespan. In addition, it is adjustable according to different QoS constraints.

Analyzing Automatically Assessed Programming Assignments In Cs1/2, Kiruthika Sivaraman

Master's Projects

This project will focus on two main objectives. The first objective is to analyze Java programming solutions submitted by students and cluster or group them based on their similarities. This will help instructors in finding repeated and unique solutions. The second objective is to analyze the assignment submission pattern of students. The metrics such as the number of times a student submits an assignment online before deadline and the amount of time a student spends on an assignment are analyzed. This information could be useful to instructors in determining the complexity of assignments and student time management.

Hiding Behind The Clouds: Efficient, Privacy-Preserving Queries Via Cloud Proxies, Surabhi Gaur

Master's Projects

This project proposes PriView, a privacy-preserving technique for querying third-party ser- vices from mobile devices. Classical private information retrieval (PIR) schemes are diffi- cult to deploy and use, since they require the target service to be replicated and modified. To avoid this problem, PriView utilizes a novel, proxy-mediated form of PIR, in which the client device fetches XORs of dummy query responses from each of two proxies and combines them to produce the required result. Unlike conventional PIR, PriView does not require the third-party service to be replicated or modified in any way. We evaluated a PriView implementation for the …

Cryptanalysis Of Homophonic Substitution-Transposition Cipher, Jeffrey Yi

Master's Projects

Homophonic substitution ciphers employ a one-to-many key to encrypt plaintext. This is in contrast to a simple substitution cipher where a one-to-one mapping is used. The advantage of a homophonic substitution cipher is that it makes frequency analysis more difficult, due to a more even distribution of plaintext statistics. Classic transposition ciphers apply diffusion to the ciphertext by swapping the order of letters. Combined transposition-substitution ciphers can be more challenging to cryptanalyze than either cipher type separately. In this research, we propose a technique to break a combined simple substitution- column transposition cipher. We also consider the related problem of …

Backward Sequential Feature Elimination And Joining Algorithms In Machine Learning, Sanya Valsan

Master's Projects

The Naïve Bayes Model is a special case of Bayesian networks with strong independence assumptions. It is typically used for classification problems. The Naïve Bayes model is trained using the given data to estimate the parameters necessary for classification. This model of classification is very popular since it is simple yet efficient and accurate. While the Naïve Bayes model is considered accurate on most of the problem instances, there is a set of problems for which the Naïve Bayes does not give accurate results when compared to other classifiers such as the decision tree algorithms. One reason for it could …

Hunting For Metamorphic Javascript Malware, Mangesh Musale

Master's Projects

Internet plays a major role in the propagation of malware. A recent trend is the infection of machines through web pages, often due to malicious code inserted in JavaScript. From the malware writer’s perspective, one potential advantage of JavaScript is that powerful code obfuscation techniques can be applied to evade de- tection. In this research, we analyze metamorphic JavaScript malware. We compare the effectiveness of several static detection strategies and we quantify the degree of morphing required to defeat each of these techniques.

Incorporating Privacy And Security Features In An Open Source Search Engine A Project Report Presented To, Akash Patel

Master's Projects

The aim of this project was to explore and implement various privacy and security features in an open-source search engine and enhance the security and privacy capabilities of Yioop. Yioop, an open-source PHP search engine based on GPLv3 license, is designed and developed by Dr. Chris Pollett. We have enabled a crawl, search and index mechanism for hidden services by execution of codes, which has facilitated access of the Tor network in Yioop. We have diversified the ability of the previously supported text CAPTCHA functionality in Yioop by implementing hash CAPTCHA and provided feasibility to toggle between text CAPTCHA and …