Malicious Software (malware) attacks across the internet are increasing at an alarming rate. Cyber-attacks have become increasingly more sophisticated and targeted. These targeted attacks are aimed at compromising networks, stealing personal financial information and removing sensitive data or disrupting operations. Current malware detection approaches work well for previously known signatures. However, malware developers utilize techniques to mutate and change software properties (signatures) to avoid and evade detection. Polymorphic malware is practically undetectable with signature-based defensive technologies. Today’s effective detection rate for polymorphic malware detection ranges from 68.75% to 81.25%. New techniques are needed to improve malware detection rates. Improved detection of polymorphic malware can only be accomplished by extracting features beyond the signature realm. Targeted detection for polymorphic malware must rely upon extracting key features and characteristics for advanced analysis. Traditionally, malware researchers have relied on limited dimensional features such as behavior (dynamic) or source/execution code analysis (static). This study’s focus was to extract and evaluate a limited set of multidimensional topological data in order to improve detection for polymorphic malware. This study used multidimensional analysis (file properties, static and dynamic analysis) with machine learning algorithms to improve malware detection. This research demonstrated improved polymorphic malware detection can be achieved with machine learning. This study conducted a number of experiments using a standard experimental testing protocol. This study utilized three advanced algorithms (Metabagging (MB), Instance Based k-Means (IBk) and Deep Learning Multi-Layer Perceptron) with a limited set of multidimensional data. Experimental results delivered detection results above 99.43%. In addition, the experiments delivered near zero false positives. The study’s approach was based on single case experimental design, a well-accepted protocol for progressive testing. The study constructed a prototype to automate feature extraction, assemble files for analysis, and analyze results through multiple clustering algorithms. The study performed an evaluation of large malware sample datasets to understand effectiveness across a wide range of malware. The study developed an integrated framework which automated feature extraction for multidimensional analysis. The feature extraction framework consisted of four modules: 1) a pre-process module that extracts and generates topological features based on static analysis of machine code and file characteristics, 2) a behavioral analysis module that extracts behavioral characteristics based on file execution (dynamic analysis), 3) an input file construction and submission module, and 4) a machine learning module that employs various advanced algorithms. As with most studies, careful attention was paid to false positive and false negative rates which reduce their overall detection accuracy and effectiveness. This study provided a novel approach to expand the malware body of knowledge and improve the detection for polymorphic malware targeting Microsoft operating systems.

Master of Science / Department of Computing and Information Sciences / William H. Hsu / The project explores the possibility of increasing efficiency in the clusters formed out of massive data sets which are formed using threshold blocking algorithm. Clusters thus formed are denser and qualitative. Clusters that are formed out of individual clustering algorithms alone, do not necessarily eliminate outliers and the clusters generated can be complex, or improperly distributed over the data set. The threshold blocking algorithm, a current research paper from Michael Higgins of Statistics Department on other hand, in comparison with existing algorithms performs better in forming the dense and distinctive units with predefined threshold. Developing a hybridized algorithm by implementing the existing clustering algorithms to re-cluster these units thus formed is part of this project. Clustering on the seeds thus formed from threshold blocking Algorithm, eases the task of clustering to the existing algorithm by eliminating the overhead of worrying about the outliers. Also, the clusters thus generated are more representative of the whole. Also, since the threshold blocking algorithm is proven to be fast and efficient, we now can predict a lot more decisions from large data sets in less time. Predicting the similar songs from Million Song Data Set using such a hybridized algorithm is considered as the data set for the evaluation of this goal.

Master of Science / Department of Computing and Information Sciences / William H. Hsu / (Higgins, Savje, & Sekhon, 2016) Provides us with a sampling blocking algorithm that enables large and complex experiments to run in polynomial time without sacrificing the precision of estimates on a covariate dataset. The goal of this project is to run the different clustering algorithms on top of clusters formed from above mentioned blocking algorithm and analyze the performance and compatibility of the clustering algorithms. We first start with applying the blocking algorithm on a covariate dataset and once the clusters are formed, we then apply our clustering algorithm HAC (Hierarchical Agglomerative Clustering) or PAM (Partitioning Around Medoids) on the seeds of the clusters. This will help us to generate more similar clusters. We compare our performance and precision of our hybridized clustering techniques with the pure clustering techniques to identify a suitable hybridized blocking model.

This thesis discusses the possibilities of using cluster analysis methods for customer segmentation. The theoretical part is focused on description of selected methods of cluster analysis and explanation of other concepts related to this topic, such as CRM, segmentation and targeted communication. In the practical part are applied cluster analysis methods to real data unnamed company with the aim of creating a default substrates useful for planning and implementation of targeted communication. For the main calculations is used program R, for data and output editing is used MS Excel. At the end of the work are evaluated applied methods and summarized lessons learned from the cluster analysis. For a company were created and characterized databases which could be useful for marketing decisions.

Intelligent Transport System (ITS) has become a hot research topic over the past decades. ITS is a system that applies the following technologies to the whole transportation management system efficiently, including information technique, wireless communication, sensor networks, control technique, and computer engineering. ITS provides an accurate, real time and synthetically efficient transportation management system. Obviously, Vehicular Ad Hoc NETworks (VANETs) attract growing attention from both the research community and industry all over the world. This is because a large amount of applications are enabled by VANETs, such as safety related applications, traffic management, commercial applications and general applications. When connecting to the internet or communicating with different networks in order to access a variety of services using VANETs, drivers and passengers in different cars need to be able to exchange messages with gateways from their vehicles. A secure gateway discovery process is therefore critical, because vehicles should not be subject to security attacks while they are communicating; however, currently there is no existing protocol focusing on secure gateway discovery. In this thesis, we first analyze and compare current existing secure service discovery protocols and then we propose a Secure Gateway Localization and Communication System for Vehicular Ad Hoc Networks (SEGAL), which concentrates on the security issue in gateway discovery. We focus on the authentication aspect by proposing secure cluster based VANETs, that can ensure the gateway discovery messages exchanged through secure clusters. We present the principle and specific process of our SEGAL protocol and analyze its performance to guarantee its outstanding practical applicability.

The main goal of this thesis is to investigate the nature of an optional consonant deletion process, through an articulatory and acoustic study of word-final consonant clusters in Persian. Persian word-final coronal stops are optionally deleted when they are preceded by obstruents or the homorganic nasal /n/. For example, the final clusters in the words /næft/ “oil”, /suχt/ “burnt” and /qæsd/ “intention” are optionally simplified in fast/casual speech, resulting in: [næf], [suχ], and [qæs]. What is not clear from this traditional description is whether the coronal stop is truly deleted, or if a coronal gesture is produced, but not heard, because it is obscured by the adjacent consonants. According to Articulatory Phonology (Browman & Goldstein 1986, 1988, 1989, 1990a, 1990b, 1992, 2001), the articulatory gestures of the deleted segments can still exist even if the segments are not heard. In this dissertation, ultrasound imaging was used to determine whether coronal consonant deletion in Persian is categorical or gradient, and the acoustic consequences of cluster simplification were investigated through duration and spectral measures. This phonetic study enables an account for the optional nature of the cluster simplification process. A general phonological account is provided for the simplification of coda clusters with rising sonority, and the acoustic and articulatory investigation focuses on the simplification of clusters with coronal stops. Ten Persian-speaking graduate students from the University of Ottawa and Carleton University, five male and five female, aged 25-38 participated in the articulatory and acoustic study. Audio and real time ultrasound video recordings were made while subjects had a guided conversation with a native speaker of Persian. 662 tokens of word-final coronal clusters were auditorily classified into unsimplified and simplified according to whether they contained an audible [t]. Singleton coda consonants and singleton /t/s were also captured as controls. The end of the constriction plateau of C1 and beginning of constriction plateau of C3 were used to define a time interval in which to measure the coronal gesture as the vertical distance between the tongue blade and the palate. Smoothing Splines ANOVA was used in a novel way to compare tongue blade height over time across the three conditions. The articulatory results of this study showed that the gestures of the deleted segments are often still present. More specifically, the findings showed that of the clusters that sounded simplified, some truly had no [t] gesture, some had gestural overlap, and some had reduced gestures. In order to explain the optional nature of the simplification process, it is argued that the simplified tokens are the result of two independent mechanisms. Inevitable mechanical and physiological effects generate gesturally reduced and overlapped tokens whereas planned language-specific behaviors driven by phonological rules or abstract cognitive representations result in no [t]-gesture output. The findings of this study support the main arguments presented in Articulatory Phonology regarding the underlying reasons for sound patterns and sound change. The results of this study are further used to examine different sound change models. It is argued that the simplified tokens with totally deleted [t] gesture could be the result of speakers changing their representations based on other people’s gestural overlap. This would be instances of the Choice and Chance categories in Blevins’ (2004) CCC sound change model. The acoustic results did not find any major cues which could distinguish simplified tokens from controls. It is argued that articulatory data should form an integral part of phonetic studies.

The aim of my thesis is to verify efficiency of different statistical methods in segmentation analysis of marketing data. I decided to carry out my study on the organic food market. The theoretical part defines segmentation from the marketing point of view, followed with statistical insight into this topic. The theory is then applied to marketing study in practical part. Finally, concrete recommendations for organic food marketers, producers and distributors are offered, which could help them to increase their revenues and to acquire a bigger market share.

This work is based on knowledge of concept of clusters and of other cooperative organizations. The general objective of this work is to solve specific, narrowly defined problems of a particular cluster. The reason was to come up with a solution that will be applicable and useful in this particular cluster. And this solution may be useful also in other clusters, because clusters throughout Europe, have many common features. Problems were generated by the author, during talks with representatives of cluster Omnipack. So these problems are important and really relevant for this cluster. Next goal was to measure these issues in relation to basic criteria, like importance, and difficulty. This evaluation was the basis for selecting problems to solve. Other objective was to search for specific solutions using existing literature, examples from other clusters and output of own research of clusters. As main problem to solve was chosen problem called "performance measurement". Because no suitable solution was found in the literature or other resources, author had to create own solution. Problematic area of the problem "performance assessment" was described as follows: "Support from European Union subsidies is really crucial for many clusters in Czech Republic right now. But after the end of these subsidies, clusters will have to prove or demonstrate to its members that they are still useful. Then cluster members will be then willing to continue with financial supporting of their cluster." The problem was solved from the perspective of cluster members, and main areas of its concept are performance evaluation of cluster members from the perspective of a cluster member, benefits of the cluster as a whole for members of the cluster, the benefits of other members of the cluster for a particular member of the cluster and cluster management role for the use of these advantages from the perspective of cluster members. Solution of this problem is very detailed and extensive, and represents main author's scientific contribution. Important point of this approach is evaluating the performance of the cluster from the perspective of cluster members. This approach brings some new insights. Author proposed his own concept of looking at the actors and factors, influencing the performance of cluster members. Proposed concept of pairwise comparison of cluster members is also really innovative. It brings new opportunities in the detailed description and understanding the cluster by its members and management.

Riders are growing more important as a part of insurance markets. The aim of this thesis is to introduce riders offered by a chosen insurance company in the Czech Republic. The next part is dedicated to the study of a structure of arranged riders in one year. A characteristic of riders and main covers a, which affect the number of arranged riders, are specified by using statistical methods. In the last part, clients are classified based on their owned riders by using cluster analysis. This thesis contains some recommenddations to create new riders.

This diploma work describes the classification of the member states of the European Union according to the demographic indicators. It evaluates development in the individual states by absolute demographic indicators too. In the year 2008 less children were born and less people were died in the most of the member states than the year 1993. The hope of the end of life grows up in all contries. Relative demographic indicators are used in the cluster analysis for diversification of rhe states into certain groups in accordance with their similarity. Two methods are used in this work --of the farthest neighbour and the Ward method (the hierarchical clustering method), the both with the Euclid distance. The hierarchical method of the farthest neighbour divided fifteen states into the four clusters in 1995, twenty-five states into the six clusters in 2004 and twenty-seven states into the six clusters in 2007. The Ward method divided these states into the three clusters in 1995, into the six clusters in 2004 and into the three clusters in 2007.