Abstract
The development in information science has enabled an explosive growth of data, which attracts more and more researchers to engage in the field of big data analytics. Noticeably, in many real-world applications, large amounts of data are imbalanced data since the events of interests occur infrequently. However, the detection of these events is such an important research problem and has attracted significant research efforts as lots of real-world big data sets have skewed class distributions. Despite extensive research efforts, rare class mining remains one of the most challenging problems in information science, especially for multimedia big data. Though inter-concept correlations have been utilized to address this issue recently, the very small number of instances in the minority class often lead to the detection of imprecise correlations and unsatisfactory classification results. This paper proposes a novel concept correlation analysis strategy framework using the correlations between the retrieval scores and labels. By integrating the correlation information, the proposed framework can help imbalance data classification and enhance rare class (or concept) mining even with trivial scores from the minority class. Experimental results on the TRECVID multimedia big benchmark data set demonstrate the effectiveness of the proposed framework with promising performance.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016 |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 281-286 |
| Number of pages | 6 |
| ISBN (Electronic) | 9781509045709 |
| DOIs | |
| State | Published - Jan 18 2017 |
| Event | 18th IEEE International Symposium on Multimedia, ISM 2016 - San Jose, United States Duration: Dec 11 2016 → Dec 13 2016 |
Other
| Other | 18th IEEE International Symposium on Multimedia, ISM 2016 |
|---|---|
| Country | United States |
| City | San Jose |
| Period | 12/11/16 → 12/13/16 |
Fingerprint
Keywords
- Concept correlation
- Imbalanced data
- Information integration
- Multimedia big data
- Rare class mining
ASJC Scopus subject areas
- Artificial Intelligence
- Computer Networks and Communications
- Media Technology
- Computer Science Applications
Cite this
Enhancing rare class mining in multimedia big data by concept correlation. / Yan, Yilin; Shyu, Mei-Ling.
Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016. Institute of Electrical and Electronics Engineers Inc., 2017. p. 281-286 7823629.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Enhancing rare class mining in multimedia big data by concept correlation
AU - Yan, Yilin
AU - Shyu, Mei-Ling
PY - 2017/1/18
Y1 - 2017/1/18
N2 - The development in information science has enabled an explosive growth of data, which attracts more and more researchers to engage in the field of big data analytics. Noticeably, in many real-world applications, large amounts of data are imbalanced data since the events of interests occur infrequently. However, the detection of these events is such an important research problem and has attracted significant research efforts as lots of real-world big data sets have skewed class distributions. Despite extensive research efforts, rare class mining remains one of the most challenging problems in information science, especially for multimedia big data. Though inter-concept correlations have been utilized to address this issue recently, the very small number of instances in the minority class often lead to the detection of imprecise correlations and unsatisfactory classification results. This paper proposes a novel concept correlation analysis strategy framework using the correlations between the retrieval scores and labels. By integrating the correlation information, the proposed framework can help imbalance data classification and enhance rare class (or concept) mining even with trivial scores from the minority class. Experimental results on the TRECVID multimedia big benchmark data set demonstrate the effectiveness of the proposed framework with promising performance.
AB - The development in information science has enabled an explosive growth of data, which attracts more and more researchers to engage in the field of big data analytics. Noticeably, in many real-world applications, large amounts of data are imbalanced data since the events of interests occur infrequently. However, the detection of these events is such an important research problem and has attracted significant research efforts as lots of real-world big data sets have skewed class distributions. Despite extensive research efforts, rare class mining remains one of the most challenging problems in information science, especially for multimedia big data. Though inter-concept correlations have been utilized to address this issue recently, the very small number of instances in the minority class often lead to the detection of imprecise correlations and unsatisfactory classification results. This paper proposes a novel concept correlation analysis strategy framework using the correlations between the retrieval scores and labels. By integrating the correlation information, the proposed framework can help imbalance data classification and enhance rare class (or concept) mining even with trivial scores from the minority class. Experimental results on the TRECVID multimedia big benchmark data set demonstrate the effectiveness of the proposed framework with promising performance.
KW - Concept correlation
KW - Imbalanced data
KW - Information integration
KW - Multimedia big data
KW - Rare class mining
UR - http://www.scopus.com/inward/record.url?scp=85015186015&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85015186015&partnerID=8YFLogxK
U2 - 10.1109/ISM.2016.75
DO - 10.1109/ISM.2016.75
M3 - Conference contribution
AN - SCOPUS:85015186015
SP - 281
EP - 286
BT - Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016
PB - Institute of Electrical and Electronics Engineers Inc.
ER -