Two of the prime open-source environments available for machine/statistical learning in data mining and knowledge discovery are the software packages Weka and R which have emerged from the machine learning and statistics communities, respectively. To make the different sets of tools from both environments available in a single unified system, an R package RWeka is suggested which interfaces Weka's functionality to R. With only a thin layer of (mostly R) code, a set of general interface generators is provided which can set up interface functions with the usual "R look and feel", re-using Weka's standardized interface of learner classes (including classifiers, clusterers, associators, filters, loaders, savers, and stemmers) with associated methods. / Series: Research Report Series / Department of Statistics and Mathematics

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

The goal of this thesis was to design the support for output breakpoints in the Java language and implement the designed solution including integration into the Eclipse IDE. An output breakpoint represents a location in an output media (e.g. file, standard output or socket) where the execution of program should stop. The location in the output media is generalized as a stop condition - it can be in the form of absolute position, a particular line and a character or an occurrence of a text string. The designed solution is based on observing the classes responsible for output using entry method breakpoints that are a standard part of the Java language. This solution is dependent on a particular version and implementation of JRE because some of observed methods are not a part of public Java API.

Approved for public release; distribution is unlimited. / The U.S. Army's Future Force is being developed as a faster, lighter, more rapidly deployable alternative to the current force structure. The Future Force will feature a smaller in-theater footprint and require the ability to cover a larger area of the battle space with intelligence-gathering assets. To support this development the Naval Postgraduate School and TRAC Monterey began to conduct research in the area of allocation of Future Force sensor platforms. A previous thesis developed the Sensor Allocation Model (SAM) for finding an appropriate mix and allocation strategy for organic Unit of Action sensors in a given threat scenario. The mix suggested by the model is robust to uncertainties in sensor performance and target quantity and location. SAM shows great promise for use as a screening tool in support of analysis of alternatives studies as well as in support of Army and Joint war fighting experimentation. It also has potential for use as an operational decision support tool for unit commanders. This thesis discusses three improvements to SAM. First, SAM has been translated into a programming language that easily can be implemented into any simulation environment. Second, it now contains more realistic constraints on sensor platform employment duration and distance. Third, the model estimates of sensor performance have been improved with a Probability Line of Sight model. Together, these improvements have greatly improved SAM's usability. / Captain, German Army

Le calcul sur grille permet à un individu, une société, de disposer d’un ensemble d’unités de calcul fournies par des tiers. Dans ce type d’architecture, des problèmes de sécurité se posent : comment garantir à la fois la sécurité du matériel et de l’application. Pour ce qui concerne la sécurité de l’application, il s’agit principalement d’assurer la confidentialité du code et l’intégrité de son exécution. La sécurité du matériel consiste à protéger la machine contre tout code malveillant pouvant être contenu dans l’application. Aucun système ne permet aujourd’hui, de satisfaire ces contraintes fortes de sécurité. Notre objectif dans cette thèse est de proposer et de valider des solutions pour les problèmes de sécurité des grilles informatiques.Pour résoudre ces problèmes, nous proposons l’utilisation de la carte à puce comme support d’exécution hautement sécurisé à travers le déploiement de grilles de cartes à puce. Grâce aux mécanismes de sécurité qu’elles intègrent, les cartes à puce permettent en effet de protéger l’application et son code aussi bien que le support d’exécution. Le travail effectué dans cette thèse consiste à définir un environnement pour le calcul sécurisé sur grille de cartes à puce. / Abstract

Over the last several years, the Web Services model of Geographic Information Systems has been rapidly evolved and materialized. In this thesis project, I have reviewed the current status of integrating the general Web Services technology (SOAP, WSDL, and UDDI) and OpenGIS Consortium (OGC) Web services standards in developing distributed GIS computing. The overlap of the web service model and the technology stack between the SOAP-based Web Services and OGC Web Services indicates the feasibility of integration. OGC has named all core general Web Services Technologies (SOAP, WSDL, UDDI) in its envisioned OWS architecture. OGC has also started putting efforts for the integration by conducting experiments, which include a SOAP experiment and an UDDI experiment. However, these experiments only identified some very specific issues based on small number of testing interfaces and scenarios. There are leading GIS software vendors who have adopted both areas in their implementation. The ESRI ArcWeb Services is a good example, which implements OGC Web Services Interfaces using SOAP, WSDL, and UDDI. In my implementation experiment, Java Web Services Developer Pack is used to build a client of Microsoft TerraService. SOAP messages are constructed to retrieve DOQ images from the TerraService as the background to display ArcSDE feature data. Query functionalities on the feature data are implemented.

The purpose of this research was to master several unfamiliar concepts and bring them together in one cohesive project. DroidCDF is an application for the Android operating system to create, write data to, and read data from netCDF format files. DroidCDF uses Unidata’s NetCDF Java Library and can write files in netCDF-3 format but read from any netCDF format files. As mobile devices become more powerful and commonplace, DroidCDF provides a convenient tool for researchers. An incremental methodology was applied; the application was built from a rough workflow to eventually a robust and fully functional program. The produced files are fully portable and can be used as the input for other applications. The application has been tested with several large netCDF files with varying conventions and has handled each one remarkably well. Upon submission of this thesis, DroidCDF will be released onto the open market.

The thesis is concerned with a comparison of two rapid web development frameworks for the Java platform - Tapestry 5 and Grails. The goal is to get acquainted with both of the frameworks, explain their basic principles and usage, compare their solution to common challenges and their contribution to faster development. Furthermore, the goal is to de fine a set of criteria for the comparison and match each one of them with their respective weights representing their importance in certain decision making scenarios. The frameworks are new and under dynamic development, therefore the bene ts of this thesis are an extensive research in foreign resources, summary of the acquired knowledge, practical usage examples and furthermore the assessment of the frameworks' strenghts and weaknesses. The conclusions and comparison quanti cation can then be used to strategically choose the appropriate framework if necessarry.

This study is conducted to show the strengths and weaknesses of C++ and Java in three areas that are used often in programming; loading, sorting and saving data.Performance and scalability are large factors in software development and choosing the right programming language is often a long process.It is important to conduct these types of direct comparison studies to properly identify strengths and weaknesses of programming languages. Two applications were created, one using C++ and one using Java.Apart from a few syntax and necessary differences, both are as close to being identical as possible.Each application loads three files containing 1000, 10000 and 100000 randomly ordered integers.These files are pre-created and always contain the same values.They are randomly generated by another small application before testing. The test runs three times, once for each file.When the data is loaded, it is sorted using quicksort.The data is reset using the dataset file and sorted again using insertion-sort.The sorted data is then saved to a file.Each test runs 50 times in a large loop and the times for loading, sorting and saving the data are saved.In total, 300 tests are run between the C++ and the Java application. The results show that Java has a total time that is faster than C++ and it is also faster when loading two out of three datasets.C++ was generally faster when sorting the datasets using both algorithms and when saving the data to files. In general Java was faster in this study, but when processing the data and when under heavy load, C++ performed better.The main difference was when loading the files.The way that Java loads the data from a file is very different from C++, even though both applications read the files character by character, Java’s “Scanner” library converts data before it parses it.With some optimization, for example by reading the file line by line and then parsing the data, C++ could be comparable or faster, but for the sake of this study, the input methods that were chosen were seemingly the fairest.

The aim of this study has been to answer the question whether it is possible to obtain notifications that work with the user, instead of against, which can be experienced as stressful and bothersome. To decrease the stressful notifications an application was created which acted as a notification control. The application used machine learning to predict when the user wanted to receive their notifications. For an artificial intelligence to work there needs to be a pattern recognition. In this case the pattern recognition that was used is called the association rule analysis. The association rule analysis used a tree called fp-growth. After the application was made, a usability test was made before and after the installation of the application. The usability test was testing if the user experienced stress and how the application worked. The study showed that screen time decreased by one hour and the number of times the mobile was opened was also reduced. This survey requires more data as it may be that the user was not affected by the application but only randomly used the mobile phone less. / Denna studie handlade om att försöka minska störande notifikationer som kan upplevas som stressande och irriterande. Det som skapades var en applikation som agerade som en notifikationskontroll. Denna applikation fungerar med hjälp av maskininlärning som ska förutse när användaren ville ta emot sina notifikationer. Den mönsterigenkännande artificiella intelligensen som användes kallas associationsregelanalys. Associationsregelanalysen använde sig av ett träd som kallas fp-growth. Det gjordes ett användartest före installation av applikationen och ett användartest efter för att se hur användaren upplevde stress men även själva applikationen. Studien visade att skärmtiden minskade med en timme och antalet gånger som mobilen öppnades minskades också. Denna undersökning kräver mer data då det kan vara så att användaren inte blev påverkad av applikationen utan endast slumpmässigt använde mobiltelefonen mindre.