Large scale tensor regression using kernels and variational inference

We outline an inherent flaw of tensor factorization models when latent factors are expressed as a function of side information and propose a novel method to mitigate this. We coin our methodology kernel fried tensor (KFT) and present it as a large-scale prediction and forecasting tool for high dimensional data. Our results show superior performance against LightGBM and Field aware factorization machines (FFM), two algorithms with proven track records, widely used in large-scale prediction. We also develop a variational inference framework for KFT which enables associating the predictions and forecasts with calibrated uncertainty estimates on several datasets.

Tensor Factorization and Clustering for the Feature Extraction Based on Tucker3 with Updating Core

Aiming at the problems of Tucker3 to large-scale tensor when applied to feature extraction, a new factorization based on Tucker3 is proposed to extract feature from the tensors. First, the large-scale tensor is divided into multiple sub-tensors so as to conveniently compute cores of sub-tensors in parallel mode with Matlab Parallel Computing Toolbox; Then, the cores of each sub-tensor are updated for reducing deviation in calculating and the similar characteristics of sub-tensors are clustered to obtain the features. Experiment results show that this methods is able to extract features rapidly and efficiently.