In a world where electronic databases are used to store ever-increasing quantities of data it is becoming harder to mine useful information from them. Therefore there is a need for a highly scalable parallel architecture capable of handling the ever-increasing complexity of data mining problems. A cycle stealing network is one possible scalable solution to this problem. A cycle stealing network allows users to donate their idle cycles to form a virtual supercomputer by connecting multiple machines via a network. This research aims to establish whether cycle stealing networks, specifically the G2 system developed at the Queensland University of Technology, are viable for large scale data mining problems. The computationally intensive sequence mining, feature selection and functional dependency mining problems are deliberately chosen to test the usefulness and scalability of G2. Tests have shown that G2 is highly scalable where the ratio of computation to communication is approximately known. However for combinatorial problems where computation times are difficult or impossible to predict, and communication costs can be unpredictable, G2 often provides little or no speedup. This research demonstrates that existing sequence mining and functional dependency mining techniques are not suited to a client-server style cycle stealing network like G2. However the feature selection is well suited to G2, and a new sequence mining algorithm offers comparable performance to other existing, non-cycle stealing, parallel sequence mining algorithms. Furthermore new functional dependency mining algorithms offer substantial benefit over existing serial algorithms.

Master of Science / Department of Computing and Information Sciences / Doina Caragea / Recent advances in social network applications have resulted in millions of users joining such networks in the last few years. User data collected from social networks can be used for various data mining problems such as interest recommendations, friendship recommendations and many more. Social networks, in general, can be seen as a huge directed network graph representing users of the network (together with their information, e.g., user interests) and their interactions (also known as friendship links). Previous work [Hsu et al., 2007] on friendship link prediction has shown that graph features contain important predictive information. Furthermore, it has been shown that user interests can be used to improve link predictions, if they are organized into an explicitly or implicitly ontology [Haridas, 2009; Parimi, 2010]. However, the above mentioned previous studies have been performed using a small set of users in the social network LiveJournal. The goal of this work is to study the performance of the ontology based approach proposed in [Haridas, 2009], when number of users in the dataset is increased. More precisely, we study the performance of the approach in terms of performance for data sets consisting of 1000, 2000, 3000 and 4000 users. Our results show that the performance generally increases with the number of users. However, the problem becomes quickly intractable from a computation time point of view. As a part of our study, we also compare our results obtained using the ontology-based approach [Haridas, 2009] with results obtained with the LDA based approach in [Parimi, 2010], when such results are available.