Wu, Q, Li, X, Do, Q, Fan, J, Ge, R & Wang, J 2018, 'CTF-PSF: Coupled Tensor Factorization with Partially Shared Factors', Proceedings of the International Joint Conference on Neural Networks, International Joint Conference on Neural Networks, IEEE, Rio de Janeiro, Brazil.View/Download from: UTS OPUS or Publisher's site
© 2018 IEEE. Coupled matrix-tensor factorization has been suc- cessfully applied in various fields in the processing of coupled data. However, the unshared components between coupled data tend to make the joint decomposition inaccurate. In order to solve this problem, in this work, we propose a method to improve the traditional method by combining individual decomposition and coupled decomposition to analyze the shared and unshared components. Numerical experiments are given to illustrate the advantages of the proposed method compared to the existing approaches.
Do, Q, Liu, W & Chen, F 2017, 'Discovering both explicit and implicit similarities for cross-domain recommendation', Advances in Knowledge Discovery and Data Mining (LNAI), Pacific Asia Conference on Advances in Knowledge Discovery and Data Mining, Springer, Jeju, South Korea, pp. 618-630.View/Download from: UTS OPUS or Publisher's site
© 2017, Springer International Publishing AG. Recommender System has become one of the most important techniques for businesses today. Improving its performance requires a thorough understanding of latent similarities among users and items. This issue is addressable given recent abundance of datasets across domains. However, the question of how to utilize this cross-domain rich information to improve recommendation performance is still an open problem. In this paper, we propose a cross-domain recommender as the first algorithm utilizing both explicit and implicit similarities between datasets across sources for performance improvement. Validated on real-world datasets, our proposed idea outperforms the current cross-domain recommendation methods by more than 2 times. Yet, the more interesting observation is that both explicit and implicit similarities between datasets help to better suggest unknown information from cross-domain sources.
Xu, Y, Gu, Y, Do, DMQ, Xie, H & Qing, L 2017, 'A Faster without Scarifying Accuracy Online Decomposition Approach for Higher-Order Tensors', Proceedings - 2017 International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery, CyberC 2017, International Conference on Cyber-Enabled Distributed Computing and Knowledge Discovery, IEEE, Nanjing, China, pp. 159-166.View/Download from: UTS OPUS or Publisher's site
© 2017 IEEE. Tensors could be very suitable for representing multidimensional data. In recent years, CANDECOMP/PARAFAC (CP) decomposition which is one of the most popular methods for Multidimensional Data Analysis has been widely studied and extensively applied. However, today's datasets will often change dynamically, and the amount of data is showing a trend of exponential growth. It is a very necessary and difficult task to perform a CP decomposition on a dynamically changing tensor with very large scale growth. The traditional and classic methods, such as Alternating Least Squares (ALS) algorithm, cannot be directly used to the dynamical tensor due to their huge consumption of time and memory. In addition, the existing online CP methods can only partially solve this problem and can only be applied to thirdorder tensor. Based on the online CP method, we proposed a simplified online CP decomposition algorithm that can be a good solution to these problems. It not only has the similar decomposition accuracy rate with ALS algorithm but also the decomposition speed faster than the ALS algorithm hundreds of thousands of times. Comparing with other state-of-theart online CP methods, it has better decomposition quality and decomposition speed. The experimental results of four methods show that, our approach reduces computational time significantly without scarifying accuracy. our approach has a similar accuracy rate, and the speed has increased by tens times than online CP decomposition. Even in some datasets, the speed and accuracy of our approach are both better than the other approach.
Do, D & Liu, W 2016, 'ASTEN: an Accurate and Scalable Approach to Coupled Tensor Factorization', the International Joint Conference in Neural Networks, The International Joint Conference in Neural Networks, IEEE, Vancouver, Canada, pp. 99-106.View/Download from: UTS OPUS or Publisher's site
Coupled Tensor Factorization (CTF) has become one of the most popular methods for joint analysis of high dimensional data generated from multiple sources. The goal of CTF is to factorize correlated datasets into latent factors efficiently. This research was taken with a particular goal of improving the accuracy of CTF. It is important to optimize the factorization of each single tensor of the coupled tensors. To achieve this, we introduce ASTEN, an Accurate and Scalable Tensor factorization method, where the objective function is optimized with respect to every single tensor and matrix. Differing from algorithms with a traditional objective function which forces shared modes among tensors to have identical factors, ASTEN enables each tensor to have its own discriminative factor on the shared mode and thus is capable of finding the accurate approximation of every tensor. Furthermore, to make it highly scalable in handling big data, we design it to be fully distributed and scalable with respect to the number of tensors, their dimensions, their sizes and the number of data partitions. In addition, we provide our theoretical proof and experimental evidence that our algorithm converges to an optimum. Experiments on both real and synthetic datasets demonstrate that our proposed ASTEN outperforms alternative existing algorithms.
Do, Q, Pham, T, Liu, W & Ramamohanarao, K 2016, 'WTEN: An Advanced Coupled Tensor Factorization Strategy for Learning from Imbalanced Data', Web Information Systems Engineering – WISE 2016, International Conference on Web Information Systems Engineering, Springer International Publishing, Shanghai, China, pp. 537-552.View/Download from: UTS OPUS or Publisher's site
Learning from imbalanced and sparse data in multi-mode and high-dimensional tensor formats efficiently is a significant problem in data mining research. On one hand, Coupled Tensor Factorization (CTF) has become one of the most popular methods for joint analysis of heterogeneous sparse data generated from different sources. On the other hand, techniques such as sampling, cost-sensitive learning, etc. have been applied to many supervised learning models to handle imbalanced data. This research focuses on studying the effectiveness of combining advantages of both CTF and imbalanced data learning techniques for missing entry prediction, especially for entries with rare class labels. Importantly, we have also investigated the implication of joint analysis of the main tensor and extra information. One of our major goals is to design a robust weighting strategy for CTF to be able to not only effectively recover missing entries but also perform well when the entries are associated with imbalanced labels. Experiments on both real and synthetic datasets show that our approach outperforms existing CTF algorithms on imbalanced data.