Computer Sciences Commons^™

Image Spam Detection, Aneri Chavda

Master's Projects

Email is one of the most common forms of digital communication. Spam can be de ned as unsolicited bulk email, while image spam includes spam text embedded inside images. Image spam is used by spammers so as to evade text-based spam lters and hence it poses a threat to email based communication. In this research, we analyze image spam detection methods based on various combinations of image processing and machine learning techniques.

Analysis Of Periodicity In Botnets, Prathiba Nagarajan

Master's Projects

A botnet consists of a network of infected computers which are controlled re- motely via a command and control (C&C) server. A typical botnet requires frequent communication between the C&C server and the infected nodes. Previous approaches to detecting botnets have employed various machine learning techniques, based on features extracted from network tra c. In this research, we carefully analyze the pe- riodicity of tra c as a means for detecting a variety of botnets by applying machine learning to publicly available datasets.

Improving Energy Efficiency For Iot Communications In 5g Networks, Gary Su

Master's Projects

Increase in number of Internet of Things (IoT) devices is quickly changing how mobile networks are being used by shifting more usage to uplink transmissions rather than downlink transmissions. Currently, mobile network uplinks utilize Single Carrier Frequency Division Multiple Access (SC-FDMA) schemes due to the low Peak to Average Power Ratio (PAPR) when compared to Orthogonal Frequency Division Multiple Access (OFDMA). In an IoT perspective, power ratios are highly important in effective battery usage since devices are typically resource-constrained. Fifth Generation (5G) mobile networks are believed to be the future standard network that will handle the influx of IoT device …

Generic Online Learning For Partial Visible & Dynamic Environment With Delayed Feedback, Behrooz Shahriari

Master's Projects

Reinforcement learning (RL) has been applied to robotics and many other domains which a system must learn in real-time and interact with a dynamic environment. In most studies the state-action space that is the key part of RL is predefined. Integration of RL with deep learning method has however taken a tremendous leap forward to solve novel challenging problems such as mastering a board game of Go. The surrounding environment to the agent may not be fully visible, the environment can change over time, and the feedbacks that agent receives for its actions can have a fluctuating delay. In this …

Mining Frequency Of Drug Side Effects Over A Large Twitter Dataset Using Apache Spark, Dennis Hsu

Master's Projects

Despite clinical trials by pharmaceutical companies as well as current FDA reporting systems, there are still drug side effects that have not been caught. To find a larger sample of reports, a possible way is to mine online social media. With its current widespread use, social media such as Twitter has given rise to massive amounts of data, which can be used as reports for drug side effects. To process these large datasets, Apache Spark has become popular for fast, distributed batch processing. In this work, we have improved on previous pipelines in sentimental analysis-based mining, processing, and extracting tweets …

Design Concept For A Failover Mechanism In Distributed Sdn Controllers, Nathan Kong

Master's Projects

Software defined networking allows the separation of the control plane and data plane in networking. It provides scalability, programmability, and centralized control. It will use these traits to reach ubiquitous connectivity. Like all concepts software defined networking does not offer these advantages without a cost. By utilizing a centralized controller, a single point of failure is created. To address this issue, this paper proposes a distributed controller failover. This failover will provide a mechanism for recovery when controllers are not located in the same location. This failover mechanism is based on number of hops from orphan nodes to the controller …

Intelligent Web Crawler For Semantic Search Engine, Shujia Zhang

Master's Projects

A Semantic Search Engine (SSE) is a program that produces semantic-oriented concepts from the Internet. A web crawler is the front end of our SSE; its primary goal is to supply important and necessary information to the data analysis component of SSE. The main function of the analysis component is to produce the concepts (moderately frequent finite sequences of keywords) from the input; it uses some variants of TF-IDF as a primary tool to remove stop words. However, it is a very expensive way to filter out stop words using the idea of TF-IDF. The goal of this project is …

Malware Detection Using The Index Of Coincidence, Bhavna Gurnani

Master's Projects

In this research, we apply the Index of Coincidence (IC) to problems in malware analysis. The IC, which is often used in cryptanalysis of classic ciphers, is a technique for measuring the repeat rate in a string of symbols. A score based on the IC is applied to a variety of challenging malware families. We nd that this relatively simple IC score performs surprisingly well, with superior results in comparison to various machine learning based scores, at least in some cases.

Cryptanalysis Of Homophonic Substitution Cipher Using Hidden Markov Models, Guannan Zhong

Master's Projects

We investigate the effectiveness of a Hidden Markov Model (HMM) with random restarts as a mean of breaking a homophonic substitution cipher. Based on extensive experiments, we find that such an HMM-based attack outperforms a previously de- veloped nested hill climb approach, particularly when the ciphertext message is short. We then consider a combination cipher, consisting of a homophonic substitution and a column transposition. We develop and analyze an attack on such a cipher. This attack employs an HMM (with random restarts), together with a hill climb to recover the column permutation. We show that this attack can succeed on …

Real-Time Online Chinese Character Recognition, Wenlong Zhang

Master's Projects

In this project, I built a web application for handwritten Chinese characters recognition in real time. This system determines a Chinese character while a user is drawing/writing it. The techniques and steps I use to build the recognition system include data preparation, preprocessing, features extraction, and classification. To increase the accuracy, two different types of neural networks ared used in the system: a multi-layer neural network and a convolutional neural network.

Handling Relationships In A Wiki System, Yashi Kamboj

Master's Projects

Wiki software enables users to manage content on the web, and create or edit web pages freely. Most wiki systems support the creation of hyperlinks on pages and have a simple text syntax for page formatting. A common, more advanced feature is to allow pages to be grouped together as categories. Currently, wiki systems support categorization of pages in a very traditional way by specifying whether a wiki page belongs to a category or not. Categorization represents unary relationship and is not sufficient to represent n-ary relationships, those involving links between multiple wiki pages.

In this project, we extend Yioop, …

Predicting User's Future Requests Using Frequent Patterns, Marc Nipuna Dominic Savio

Master's Projects

In this research, we predict User's Future Request using Data Mining Algorithm. Usage of the World Wide Web has resulted in a huge amount of data and handling of this data is getting hard day by day. All this data is stored as Web Logs and each web log is stored in a different format with different Field names like search string, URL with its corresponding timestamp, User ID’s that helps for session identification, Status code, etc. Whenever a user requests for a URL there is a delay in getting the page requested and sometimes the request is denied. Our …

Deep Data Analysis On The Web, Xuanyu Liu

Master's Projects

Search engines are well known to people all over the world. People prefer to use keywords searching to open websites or retrieve information rather than type typical URLs. Therefore, collecting finite sequences of keywords that represent important concepts within a set of authors is important, in other words, we need knowledge mining. We use a simplicial concept method to speed up concept mining. Previous CS 298 project has studied this approach under Dr. Lin. This method is very fast, for example, to mine the concept, FP-growth takes 876 seconds from a database with 1257 columns 65k rows, simplicial complex only …

Web-Based Integrated Development Environment, Hien T. Vu

Master's Projects

As tablets become more powerful and more economical, students are attracted to them and are moving away from desktops and laptops. Their compact size and easy to use Graphical User Interface (GUI) reduce the learning and adoption barriers for new users. This also changes the environment in which undergraduate Computer Science students learn how to program. Popular Integrated Development Environments (IDE) such as Eclipse and NetBeans require disk space for local installations as well as an external compiler. These requirements cannot be met by current tablets and thus drive the need for a web-based IDE. There are also many other …

Analyzing Clustered Web Concepts With Homology, Eric Nam

Master's Projects

As data is being mined more and more from the Internet today, Data Science has become an important field of computing to make that data useful. Data Science allows people to turn all of that data into structured knowledge that is easily utilized, validated, and understandable. There are many known theories to analyze data, but this project will focus on a recently introduced method: analyzing text data with homology from mathematics to understand relationships between keyword-sets.

Using structures of algebraic topology as a starting point, keyword-sets in the text are represented by simplexes based on what they are and what …

Analysis On Alergia Algorithm: Pattern Recognition By Automata Theory, Xuanyi Qi

Master's Projects

Based on Kolmogorov Complexity, a finite set x of strings has a pattern if the set x can be output by a Turing machine of length that is less than minimum of all |x|; this Turing machine, that may not be unique, is called a pattern of the finite set of string. In order to find a pattern of a given finite set of strings (assuming such a pattern exists), the ALERGIA algorithm is used to approximate such a pattern (Turing machine) in terms of finite automata. Note that each finite automaton defines a partition on formal language Σ*, ALERGIA …

Analyze Large Multidimensional Datasets Using Algebraic Topology, David Le

Master's Projects

This paper presents an efficient algorithm to extract knowledge from high-dimensionality, high- complexity datasets using algebraic topology, namely simplicial complexes. Based on concept of isomorphism of relations, our method turn a relational table into a geometric object (a simplicial complex is a polyhedron). So, conceptually association rule searching is turned into a geometric traversal problem. By leveraging on the core concepts behind Simplicial Complex, we use a new technique (in computer science) that improves the performance over existing methods and uses far less memory. It was designed and developed with a strong emphasis on scalability, reliability, and extensibility. This paper …

Dna Analysis Using Grammatical Inference, Cory Cook

Master's Projects

An accurate language definition capable of distinguishing between coding and non-coding DNA has important applications and analytical significance to the field of computational biology. The method proposed here uses positive sample grammatical inference and statistical information to infer languages for coding DNA.

An algorithm is proposed for the searching of an optimal subset of input sequences for the inference of regular grammars by optimizing a relevant accuracy metric. The algorithm does not guarantee the finding of the optimal subset; however, testing shows improvement in accuracy and performance over the basis algorithm.

Testing shows that the accuracy of inferred languages for …

Image Spam Analysis, Annapurna Sowmya Annadatha

Master's Projects

Image spam is unsolicited bulk email, where the message is embedded in an image. This technique is used to evade text-based spam lters. In this research, we analyze and compare two novel approaches for detecting spam images. Our rst approach focuses on the extraction of a broad set of image features and selection of an optimal subset using a Support Vector Machine (SVM). Our second approach is based on Principal Component Analysis (PCA), where we determine eigenvectors for a set of spam images and compute scores by projecting images onto the resulting eigenspace. Both approaches provide high accuracy with low …

Defeating N-Gram Scores For Http Attack Detection, Samyuktha Sridharan

Master's Projects

Web applications that generate malicious HTTP requests provide a platform that attackers use to exploit vulnerable machines. Such malicious traffic should be identified by network intrusion detection systems, based on traffic analysis. Previous research has shown that n-gram techniques can be successfully applied to detect HTTP attacks. In this research, we analyze the robustness of these n-gram techniques. We show that n-gram scores are surprisingly robust, but can be defeated using certain obfuscation strategies. We also consider the need for a more costlier HMM-based intrusion detection system.

Machine Learning On The Cloud For Pattern Recognition, Tien Nguyen

Master's Projects

Pattern recognition is a field of machine learning with applications to areas such as text recognition and computer vision. Machine learning algorithms, such as convolutional neural networks, may be trained to classify images. However, such tasks may be computationally intensive for a commercial computer for larger volumes or larger sizes of images. Cloud computing allows one to overcome the processing and memory constraints of average commercial computers, allowing computations on larger amounts of data. In this project, we developed a system for detection and tracking of moving human and vehicle objects in videos in real time or near real time. …

Static And Dynamic Analysis For Android Malware Detection, Ankita Kapratwar

Master's Projects

Static analysis relies on features extracted without executing code, while dynamic analysis extracts features based on code execution (or emulation). In general, static analysis is more e cient, while static analysis is often more informative, particularly in cases of highly obfuscated code. Static analysis of an Android application can rely on features extracted from the manifest le or the Java bytecode, while dynamic analysis of Android applications can deal with features involving dynamic code loading and system calls that are collected while the application is running. In this research, we analyzed the e ectiveness of combining static and dynamic features …

VigenèRe Score For Malware Detection, Suchita Deshmukh

Master's Projects

Previous research has applied classic cryptanalytic techniques to the malware detection problem. Speci cally, scores based on simple substitution cipher cryptanal- ysis and various generalizations have been considered. In this research, we analyze two new malware scoring techniques based on classic cryptanalysis. Our rst ap- proach relies on the Index of Coincidence, which is used, for example, to determine the length of the keyword in a Vigenère ciphertext. We also consider a score based on a more complete cryptanalysis of a Vigenère cipher. We nd that the Vigenère score is competitive with previous statistical-based malware scores.

Supervised Learning For Multi-Domain Text Classification, Siva Charan Reddy Gangireddy

Master's Projects

Digital information available on the Internet is increasing day by day. As a result of this, the demand for tools that help people in finding and analyzing all these resources are also growing in number. Text Classification, in particular, has been very useful in managing the information. Text Classification is the process of assigning natural language text to one or more categories based on the content. It has many important applications in the real world. For example, finding the sentiment of the reviews, posted by people on restaurants, movies and other such things are all applications of Text classification. In …

Multi Faceted Text Classification Using Supervised Machine Learning Models, Abhiteja Gajjala

Master's Projects

In recent year’s document management tasks (known as information retrieval) increased a lot due to availability of digital documents everywhere. The need of automatic methods for extracting document information became a prominent method for organizing information 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. In my research classification of text is mainly focused on sentiment label classification. The idea proposed for sentiment analysis is multi-class classification of online movie reviews. Many research papers discussed the classification of sentiment either positive or …

Hybrid Similarity Function For Big Data Entity Matching With R-Swoosh, Vimal Chandra Gorijala

Master's Projects

Entity Matching (EM) is the problem of determining if two entities in a data set refer to the same real-world object. For example, it decides if two given mentions in the data, such as “Helen Hunt” and “H. M. Hunt”, refer to the same real-world entity by using different similarity functions. This problem plays a key role in information integration, natural language understanding, information processing on the World-Wide Web, and on the emerging Semantic Web. This project deals with the similarity functions and thresholds utilized in them to determine the similarity of the entities. The work contains two major parts: …

Efficient Pair-Wise Similarity Computation Using Apache Spark, Parineetha Gandhi Tirumali

Master's Projects

Entity matching is the process of identifying different manifestations of the same real world entity. These entities can be referred to as objects(string) or data instances. These entities are in turn split over several databases or clusters based on the signatures of the entities. When entity matching algorithms are performed on these databases or clusters, there is a high possibility that a particular entity pair is compared more than once. The number of comparison for any two entities depend on the number of common signatures or keys they possess. This effects the performance of any entity matching algorithm. This paper …

Library Writers Reward Project, Saravana Kumar Gajendran

Master's Projects

Open-source library development exploits the distributed intelligence of participants in Internet communities. Nowadays, contribution to the open-source community is fading [16] (Stackalytics, 2016) as there is not much recognition for library writers. They can start exploring ways to generate revenue as they actively contribute to the open-source community.

This project helps library writers to generate revenue in the form of bitcoins for their contribution. Our solution to generate revenue for library writers is to integrate bitcoin mining with existing JavaScript libraries, such as jQuery. More use of the library leads to more revenue for the library writers. It uses the …

Movie Script Shot Lister, David Robert Smith

Master's Projects

The making of a motion picture almost always starts with the script, the written version of a story envisioned within the mind of its creator. The script is then broken down into shots. Each individual shot is filmed and then they are edited together to create the motion picture. The goal of the Movie Script Shot Lister thesis project is to be able to read in a script for a movie or television show, and automatically generate a shot list. While a script is text, a shot list is the blue print for how to visualize that script, so the …

Malicious Javascript Detection Using Statistical Language Model, Anumeha Shah

Master's Projects

The Internet has an immense importance in our day to day life, but at the same time, it has become the medium of infecting computers, attacking users, and distributing malicious code. As JavaScript is the principal language of client side pro- gramming, it is frequently used in conducting such attacks. Various approaches have been made to overcome the JavaScript security issues. Some advanced approaches utilize machine learning technology in combination with de-obfuscation and emula- tion. Many methods of analysis incorporate static analysis and dynamic analysis. Our solution is entirely based on static analysis, which avoids unnecessary runtime overhead.

The central …