Atjaunināt sīkdatņu piekrišanu

Matrix and Tensor Factorization Techniques for Recommender Systems 1st ed. 2016 [Mīkstie vāki]

,
  • Formāts: Paperback / softback, 102 pages, height x width: 235x155 mm, weight: 1766 g, 22 Illustrations, color; 29 Illustrations, black and white; VI, 102 p. 51 illus., 22 illus. in color., 1 Paperback / softback
  • Sērija : SpringerBriefs in Computer Science
  • Izdošanas datums: 06-Feb-2017
  • Izdevniecība: Springer International Publishing AG
  • ISBN-10: 3319413562
  • ISBN-13: 9783319413563
Citas grāmatas par šo tēmu:
  • Mīkstie vāki
  • Cena: 69,22 €*
  • * ši ir gala cena, t.i., netiek piemērotas nekādas papildus atlaides
  • Standarta cena: 81,44 €
  • Ietaupiet 15%
  • Grāmatu piegādes laiks ir 3-4 nedēļas, ja grāmata ir uz vietas izdevniecības noliktavā. Ja izdevējam nepieciešams publicēt jaunu tirāžu, grāmatas piegāde var aizkavēties.
  • Daudzums:
  • Ielikt grozā
  • Piegādes laiks - 4-6 nedēļas
  • Pievienot vēlmju sarakstam
Matrix and Tensor Factorization Techniques for Recommender Systems 1st ed. 2016Palielināt
  • Formāts: Paperback / softback, 102 pages, height x width: 235x155 mm, weight: 1766 g, 22 Illustrations, color; 29 Illustrations, black and white; VI, 102 p. 51 illus., 22 illus. in color., 1 Paperback / softback
  • Sērija : SpringerBriefs in Computer Science
  • Izdošanas datums: 06-Feb-2017
  • Izdevniecība: Springer International Publishing AG
  • ISBN-10: 3319413562
  • ISBN-13: 9783319413563
Citas grāmatas par šo tēmu:
Permanent link: https://www.kriso.lv/db/9783319413563.html
Keywords:
This book presents the algorithms used to provide recommendations by exploiting matrix factorization and tensor decomposition techniques. It highlights well-known decomposition methods for recommender systems, such as Singular Value Decomposition (SVD), UV-decomposition, Non-negative Matrix Factorization (NMF), etc. and describes in detail the pros and cons of each method for matrices and tensors. This book provides a detailed theoretical mathematical background of matrix/tensor factorization techniques and a step-by-step analysis of each method on the basis of an integrated toy example that runs throughout all its chapters and helps the reader to understand the key differences among methods. It also contains two chapters, where different matrix and tensor methods are compared experimentally on real data sets, such as Epinions, GeoSocialRec, Last.fm, BibSonomy, etc. and provides further insights into the advantages and disadvantages of each method.The book offers a rich blend of

theory and practice, making it suitable for students, researchers and practitioners interested in both recommenders and factorization methods. Lecturers can also use it for classes on data mining, recommender systems and dimensionality reduction methods.

Part I Matrix Factorization Techniques.- 1. Introduction.- 2. Related Work on Matrix Factorization.- 3. Performing SVD on matrices and its Extensions.- 4. Experimental Evaluation on Matrix Decomposition Methods.- Part II Tensor Factorization Techniques.- 5. Related Work on Tensor Factorization.- 6. HOSVD on Tensors and its Extensions.- 7. Experimental Evaluation on Tensor Decomposition Methods.- 8 Conclusions and Future Work.

Recenzijas

This carefully written book offers advanced undergraduates, graduate students, researchers and professionals a comprehensive overview of the general concepts and techniques (e.g., models and algorithms) related to matrix and tensor factorization in the field of recommender systems, with a rich blend of theory and practice. I am definitely a recommender of this book! (Bruno Carpentieri, Mathematical Reviews, August, 2017)

Part I Matrix Factorization Techniques
1 Introduction
3(16)
1.1 Recommender Systems
3(2)
1.2 Recommender Systems in Social Media
5(1)
1.3 Matrix Factorization
6(2)
1.4 Tensor Factorization
8(2)
1.5 Mathematical Background and Notation
10(3)
1.6 Book Outline
13(6)
References
14(5)
2 Related Work on Matrix Factorization
19(14)
2.1 Dimensionality Reduction on Matrices
19(1)
2.2 Eigenvalue Decomposition
20(2)
2.3 Nonnegative Matrix Factorization
22(2)
2.4 Latent Semantic Indexing
24(1)
2.5 Probabilistic Latent Semantic Indexing
25(1)
2.6 CUR Matrix Decomposition
26(3)
2.7 Other Matrix Decomposition Methods
29(4)
References
30(3)
3 Performing SVD on Matrices and Its Extensions
33(26)
3.1 Singular Value Decomposition (SVD)
33(7)
3.1.1 Applying the SVD and Preserving the Largest Singular Values
37(1)
3.1.2 Generating the Neighborhood of Users/Items
38(1)
3.1.3 Generating the Recommendation List
39(1)
3.1.4 Inserting a Test User in the c-Dimensional Space
39(1)
3.2 From SVD to UV Decomposition
40(19)
3.2.1 Objective Function Formulation
43(1)
3.2.2 Avoiding Overfitting with Regularization
44(1)
3.2.3 Incremental Computation of UV Decomposition
45(5)
3.2.4 The UV Decomposition Algorithm
50(1)
3.2.5 Fusing Friendship into the Objective Function
51(2)
3.2.6 Inserting New Data in the Initial Matrix
53(4)
References
57(2)
4 Experimental Evaluation on Matrix Decomposition Methods
59(10)
4.1 Data Sets
59(1)
4.2 Sensitivity Analysis of the UV Decomposition Algorithm
60(3)
4.2.1 Tuning of the k Latent Feature Space
60(1)
4.2.2 Tuning of Parameter β
61(1)
4.2.3 Optimizing Algorithm's Parameters
62(1)
4.3 Comparison to Other Decomposition Methods
63(6)
References
65(4)
Part II Tensor Factorization Techniques
5 Related Work on Tensor Factorization
69(12)
5.1 Preliminary Knowledge of Tensors
69(3)
5.2 Tucker Decomposition and HOSVD
72(1)
5.3 AlsHOSVD
73(1)
5.4 Parallel Factor Analysis (PARAFAC)
74(1)
5.5 Pairwise Interaction Tensor Factorization (PITF)
75(1)
5.6 PCLAF and RPCLAF Algorithms
76(2)
5.7 Other Tensor Decomposition Methods
78(3)
References
79(2)
6 HOSVD on Tensors and Its Extensions
81(14)
6.1 Algorithm's Outline
81(1)
6.2 HOSVD in STSs
82(7)
6.2.1 Handling the Sparsity Problem
85(1)
6.2.2 Inserting New Users, Tags, or Items
85(1)
6.2.3 Update by Folding-in
86(2)
6.2.4 Update by Incremental SVD
88(1)
6.3 Limitations and Extensions of HOSVD
89(6)
6.3.1 Combining HOSVD with a Content-Based Method
90(1)
6.3.2 Combining HOSVD with a Clustering Method
91(2)
References
93(2)
7 Experimental Evaluation on Tensor Decomposition Methods
95(6)
7.1 Data Sets
95(1)
7.2 Experimental Protocol and Evaluation Metrics
96(1)
7.3 Sensitivity Analysis of the HOSVD Algorithm
96(2)
7.4 Comparison of HOSVD with Other Tensor Decomposition Methods in STSs
98(3)
References
99(2)
8 Conclusions and Future Work
101
Panagiotis Symeonidis is Adjunct Assistant Professor at the Aristotle University of Thessaloniki, Greece. He is the co-author of 2 international books, 18 journal papers, 4 book chapters and more than 28 articles in international conference proceedings. His articles have received almost 1400 citations from other scientific publications. He teaches courses on databases, data mining and data. For almost four years, he was the head of 1st EK (Laboratory Center) of Stavroupolis between September 2011 to July 2015. His research interests focus on recommender systems, social media in Web 2.0 and time-evolving online social networks. Andreas Zioupos has a B.Sc. degree in Mathematics and received his M.Sc. degree in Informatics & Management in 2015 from the Aristotle University of Thessaloniki, under the supervision of Dr. Panagiotis Symeonidis. He is an instructor for Google web tools and also has currently a contract as freelancer with the University of Piraeus on the project Creating a framework for documentation, collection and disposal in the form of Linked Open Data from research results and official data of general government relating to domestic economic activity. His research interests focus on data mining, recommender systems and dimensionality reduction methods.