Abstract
We propose the approximation-theoretic technique of optimal recovery for imputing missing values in clustered data, specifically for non-negative matrix factorization (NMF), and develop an algorithm for implementation. Under certain geometric conditions, we prove tight upper bounds on NMF relative error, which is the first bound of this type for missing values. Experiments on image data and biological data show that this technique performs as well as or better than other imputation techniques that account for local structure.
| Original language | English (US) |
|---|---|
| Title of host publication | 2019 IEEE Data Science Workshop, DSW 2019 - Proceedings |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 180-184 |
| Number of pages | 5 |
| ISBN (Electronic) | 9781728107080 |
| DOIs | |
| State | Published - Jun 2019 |
| Event | 2019 IEEE Data Science Workshop, DSW 2019 - Minneapolis, United States Duration: Jun 2 2019 → Jun 5 2019 |
Publication series
| Name | 2019 IEEE Data Science Workshop, DSW 2019 - Proceedings |
|---|
Conference
| Conference | 2019 IEEE Data Science Workshop, DSW 2019 |
|---|---|
| Country | United States |
| City | Minneapolis |
| Period | 6/2/19 → 6/5/19 |
Fingerprint
Keywords
- imputation
- missing values
- non-negative matrix factorization
- optimal recovery
ASJC Scopus subject areas
- Computer Networks and Communications
- Safety, Risk, Reliability and Quality
- Computational Theory and Mathematics
- Artificial Intelligence
Cite this
Non-Negative Matrix Factorization of Clustered Data with Missing Values. / Chen, Rebecca; Varshney, Lav R.
2019 IEEE Data Science Workshop, DSW 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 180-184 8755555 (2019 IEEE Data Science Workshop, DSW 2019 - Proceedings).Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Non-Negative Matrix Factorization of Clustered Data with Missing Values
AU - Chen, Rebecca
AU - Varshney, Lav R
PY - 2019/6
Y1 - 2019/6
N2 - We propose the approximation-theoretic technique of optimal recovery for imputing missing values in clustered data, specifically for non-negative matrix factorization (NMF), and develop an algorithm for implementation. Under certain geometric conditions, we prove tight upper bounds on NMF relative error, which is the first bound of this type for missing values. Experiments on image data and biological data show that this technique performs as well as or better than other imputation techniques that account for local structure.
AB - We propose the approximation-theoretic technique of optimal recovery for imputing missing values in clustered data, specifically for non-negative matrix factorization (NMF), and develop an algorithm for implementation. Under certain geometric conditions, we prove tight upper bounds on NMF relative error, which is the first bound of this type for missing values. Experiments on image data and biological data show that this technique performs as well as or better than other imputation techniques that account for local structure.
KW - imputation
KW - missing values
KW - non-negative matrix factorization
KW - optimal recovery
UR - http://www.scopus.com/inward/record.url?scp=85069536911&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85069536911&partnerID=8YFLogxK
U2 - 10.1109/DSW.2019.8755555
DO - 10.1109/DSW.2019.8755555
M3 - Conference contribution
AN - SCOPUS:85069536911
T3 - 2019 IEEE Data Science Workshop, DSW 2019 - Proceedings
SP - 180
EP - 184
BT - 2019 IEEE Data Science Workshop, DSW 2019 - Proceedings
PB - Institute of Electrical and Electronics Engineers Inc.
ER -