haku: @keyword MATLAB / yhteensä: 36
viite: 13 / 36
Tekijä: | Wang, Chao |
Työn nimi: | Software development of quadratic nonnegative matrix factorization |
Julkaisutyyppi: | Diplomityö |
Julkaisuvuosi: | 2011 |
Sivut: | x + 50 s. + liitt. 11 Kieli: eng |
Koulu/Laitos/Osasto: | Tietotekniikan laitos |
Oppiaine: | Informaatiotekniikka (T-61) |
Valvoja: | Oja, Erkki |
Ohjaaja: | Yang, Zhirong |
OEVS: | Sähköinen arkistokappale on luettavissa Aalto Thesis Databasen kautta.
Ohje Digitaalisten opinnäytteiden lukeminen Aalto-yliopiston Harald Herlin -oppimiskeskuksen suljetussa verkossaOppimiskeskuksen suljetussa verkossa voi lukea sellaisia digitaalisia ja digitoituja opinnäytteitä, joille ei ole saatu julkaisulupaa avoimessa verkossa. Oppimiskeskuksen yhteystiedot ja aukioloajat: https://learningcentre.aalto.fi/fi/harald-herlin-oppimiskeskus/ Opinnäytteitä voi lukea Oppimiskeskuksen asiakaskoneilla, joita löytyy kaikista kerroksista.
Kirjautuminen asiakaskoneille
Opinnäytteen avaaminen
Opinnäytteen lukeminen
Opinnäytteen tulostus
|
Sijainti: | P1 Ark Aalto 7128 | Arkisto |
Avainsanat: | quadratic nonnegative matrix factorization auxiliary function multiplicative update rules matlab toolbox |
Tiivistelmä (eng): | Nonnegative Matrix Factorization (NMF) has become an increasingly important research field during the past few years and has shown advantages in clustering and learning part-based representations, etc. Recently Yang and Oja have systematically studied a variant of NMF called Quadratic Nonnegative Matrix Factorization (QNMF) where factorizing matrices may appear twice in the approximation. The multiplicative update rules of QNMF with different divergence types have the same forms but different parameter values. In real-world applications, it could be tedious if the user wants to implement and compare different types of QNMF with various distance measurements and factorizing forms. In this thesis work, a QNMF toolbox is developed, which covers the implementations of diverse forms of QNMF, such as Projective Nonnegative Matrix Factorization (PNMF), Asymmetric QNMF (AQNMF), Symmetric QNMF (SNMF), Symmetric 3-Factor NMF. Various divergence types that calculate the distance between the input matrix and its approximation are also included in our toolbox. Furthermore, we present how to derive the multiplicative update rules of linear and quadratic factorizing matrices in QNMF, especially how to build an auxiliary function. Finally, we have conducted three experiments on both synthetic and real-world datasets to testify the advantages of QNMF and the toolbox in the applications of feature extraction, biclustering and estimating Hidden Markov Models for genetic sequence, compared with those implementations without using our QNMF toolbox. Experimental results show that QNMF is effective for these applications. The QNMF toolbox is very flexible to be integrated in the user's own applications and significantly facilitates the implementations and comparisons of different QNMF variants. |
ED: | 2011-08-16 |
INSSI tietueen numero: 42645
+ lisää koriin
INSSI