Paper Abstract and Keywords |
Presentation |
2020-01-30 10:00
Construction of Basis Vectors for Representation of Immunostaining Combination by Non-negative Matrix Decomposition Kaho Ko, Noriaki Hashimoto, Tatsuya Yokota (NITech), Masato Nakaguro, Kei Kohno, Shigeo Nakamura (NUH), Ichiro Takeuchi, Hidekata Hontani (NITech) MI2019-99 |
Abstract |
(in Japanese) |
(See Japanese page) |
(in English) |
In this paper, we propose a method that constructs a set of basis vectors for representing combination of immunostaining used to diagnose malignant lymphoma with fewer parameters by means of constrained non-negative matrix factorization. Immunostaining combinations can be represented by binary matrices, of which operations are Boolean, and it is known that matrix decomposition with Boolean operation is NP-hard. We hence relax this problem to a constrained non-negative real matrix decomposition and solve the relaxed problem using an ADMM (Alternating Direction Method of Multipliers). In addition, we introduce an L1-regularization technique to make the method stable against stains infrequently used. We show the concrete algorithms and demonstrate the experimental results that show the performance of the proposed method. |
Keyword |
(in Japanese) |
(See Japanese page) |
(in English) |
Non-negative matrix decomposition / Boolean Matrix Factorization / Text analysis / / / / / |
Reference Info. |
IEICE Tech. Rep., vol. 119, no. 399, MI2019-99, pp. 151-154, Jan. 2020. |
Paper # |
MI2019-99 |
Date of Issue |
2020-01-22 (MI) |
ISSN |
Online edition: ISSN 2432-6380 |
Copyright and reproduction |
All rights are reserved and no part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publisher. Notwithstanding, instructors are permitted to photocopy isolated articles for noncommercial classroom use without fee. (License No.: 10GA0019/12GB0052/13GB0056/17GB0034/18GB0034) |
Download PDF |
MI2019-99 |
|