Abstract | ||
---|---|---|
•We describe application of conformal prediction to matrix factorization to provide confidence values to each of the predictions made by the matrix factorization algorithm.•We propose four different ways of applying conformal prediction to matrix factorization.•We define different nonconformity measures suitable to matrix factoriza-tion.•We experimentally proved the best nonconformity measure among the proposed nonconformity measures.•We empirically demonstrate validity and efficiency of our CP algorithm and why our proposed algorithm failed in producing the tight prediction regions at higher confidence levels.•We experimentally demonstrate the best approach between TCP and ICP for matrix factorization problem.•We also compare execution time of the proposed methods with matrix factorization. |
Year | DOI | Venue |
---|---|---|
2018 | 10.1016/j.ins.2018.04.004 | Information Sciences |
Keywords | Field | DocType |
Recommender systems,Conformal prediction,Matrix factorization,Confidence,Nonconformity measure | Nonconformity,Recommender system,Matrix (mathematics),Matrix decomposition,Word error rate,Conformal map,Artificial intelligence,Confidence interval,Mathematics,Machine learning | Journal |
Volume | ISSN | Citations |
467 | 0020-0255 | 10 |
PageRank | References | Authors |
0.42 | 25 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Tadiparthi V. R. Himabindu | 1 | 11 | 1.11 |
Vineet Padmanabhan | 2 | 216 | 25.90 |
Arun K. Pujari | 3 | 420 | 48.20 |