# RDFTensor v1.3

0

0th

Percentile

## Different Tensor Factorization (Decomposition) Techniques for RDF Tensors (Three-Mode-Tensors)

Different Tensor Factorization techniques suitable for RDF Tensors. RDF Tensors are three-mode-tensors, binary tensors and usually very sparse. Currently implemented methods are 'RESCAL' Maximilian Nickel, Volker Tresp, and Hans-Peter Kriegel (2012) <doi:10.1145/2187836.2187874>, 'NMU' Daniel D. Lee and H. Sebastian Seung (1999) <doi:10.1038/44565>, 'ALS', Alternating Least Squares 'parCube' Papalexakis, Evangelos, C. Faloutsos, and N. Sidiropoulos (2012) <doi:10.1007/978-3-642-33460-3_39>, 'CP_APR' C. Chi and T. G. Kolda (2012) <doi:10.1137/110859063>. The code is mostly converted from MATLAB and Python implementations of these methods. The package also contains functions to get Boolean (Binary) transformation of the real-number-decompositions. These methods also are for general tensors, so with few modifications they can be applied for other types of tensor.

## Functions in RDFTensor

 Name Description default_parcube_options parCube initialization parameters CP_01 transformation of the real-number-CP-decompositions to Binary getTensor getting tensor from triples CP_R01 inverse of real-number-CP-decompositions to Binary cp_nmu Compute nonnegative CP with multiplicative updates(NMU) RDFTensor-package RDFTensor Tensor_error RESCAL Tensor error cp_als Compute a CP decomposition using an alternating least-squares algorithm(als) RescalReconstructBack reconstruct a tensor from RESCAL factorization result getTensor3m getting tensor from triples cp_apr Compute nonnegative CP with alternating Poisson regression(CP_APR) rescal_SO_Val calaculate scores Subj,Predicate,Obj serial_parCube Serial 3-mode ParCube algorithm for memory resident tensors spt_mttkrp Matricized tensor times Khatri-Rao product for ktensor tensor2mat Reduce tensor to matrix getTnsrijk tensor frontal slices to indices scRescal Scalable RESCAL Factorization rescal_Trp_Val RESCAL scores of triples inv_rescal_sf_prd_chnkgrp Reconstruct predicate(s) via chunks and grouping recon_1prd_topn_par Reconstruct predicate top scores on very large graphs rescal RESCAL: Tensor Factorization. tnsr2trp sparse tensor to triples rescal_01 transformation of the real-number-RESCAL-decompositions to Binary umls_tnsr RDF tensor of UMLS small graph No Results!