OcVFDT: One-class very fast decision tree for one-class classification of data streams

Chen Li; Yang Zhang; Xue Li

doi:10.1145/1601966.1601981

OcVFDT: One-class very fast decision tree for one-class classification of data streams

Chen Li, Yang Zhang, Xue Li

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

48 Citations (Scopus)

Abstract

Current research on data stream classification mainly focuses on supervised learning, in which a fully labeled data stream is needed for training. However, fully labeled data streams are expensive to obtain, which make the supervised learning approach difficult to be applied to real-life applications. In this paper, we model applications, such as credit fraud detection and intrusion detection, as a one-class data stream classification problem. The cost of fully labeling the data stream is reduced as users only need to provide some positive samples together with the unlabeled samples to the learner. Based on VFDT and POSC4.5, we propose our OcVFDT (One-class Very Fast Decision Tree) algorithm. Experimental study on both synthetic and real-life datasets shows that the OcVFDT has excellent classification performance. Even 80% of the samples in data stream are unlabeled, the classification performance of OcVFDT is still very close to that of VFDT, which is trained on fully labeled stream.

Original language	English
Title of host publication	Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09
Publisher	Association for Computing Machinery (ACM)
Pages	79-86
Number of pages	8
ISBN (Print)	9781605586687
DOIs	https://doi.org/10.1145/1601966.1601981
Publication status	Published - 2009
Externally published	Yes
Event	3rd International Workshop on Knowledge Discovery from Sensor Data - Paris, France Duration: 28 Jun 2009 → 28 Jun 2009 Conference number: 3

Conference

Conference	3rd International Workshop on Knowledge Discovery from Sensor Data
Country/Territory	France
City	Paris
Period	28/06/09 → 28/06/09
Other	Held in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09

Access to Document

10.1145/1601966.1601981

Cite this

Li, C., Zhang, Y., & Li, X. (2009). OcVFDT: One-class very fast decision tree for one-class classification of data streams. In Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09 (pp. 79-86). [1601981] Association for Computing Machinery (ACM). https://doi.org/10.1145/1601966.1601981

Li, Chen ; Zhang, Yang ; Li, Xue. / OcVFDT : One-class very fast decision tree for one-class classification of data streams. Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09. Association for Computing Machinery (ACM), 2009. pp. 79-86

@inproceedings{e2d942c9b054492abb978c0176d264a5,

title = "OcVFDT: One-class very fast decision tree for one-class classification of data streams",

abstract = "Current research on data stream classification mainly focuses on supervised learning, in which a fully labeled data stream is needed for training. However, fully labeled data streams are expensive to obtain, which make the supervised learning approach difficult to be applied to real-life applications. In this paper, we model applications, such as credit fraud detection and intrusion detection, as a one-class data stream classification problem. The cost of fully labeling the data stream is reduced as users only need to provide some positive samples together with the unlabeled samples to the learner. Based on VFDT and POSC4.5, we propose our OcVFDT (One-class Very Fast Decision Tree) algorithm. Experimental study on both synthetic and real-life datasets shows that the OcVFDT has excellent classification performance. Even 80% of the samples in data stream are unlabeled, the classification performance of OcVFDT is still very close to that of VFDT, which is trained on fully labeled stream.",

author = "Chen Li and Yang Zhang and Xue Li",

year = "2009",

doi = "10.1145/1601966.1601981",

language = "English",

isbn = "9781605586687",

pages = "79--86",

booktitle = "Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09",

publisher = "Association for Computing Machinery (ACM)",

address = "United States of America",

note = "3rd International Workshop on Knowledge Discovery from Sensor Data ; Conference date: 28-06-2009 Through 28-06-2009",

}

Li, C, Zhang, Y & Li, X 2009, OcVFDT: One-class very fast decision tree for one-class classification of data streams. in Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09., 1601981, Association for Computing Machinery (ACM), pp. 79-86, 3rd International Workshop on Knowledge Discovery from Sensor Data, Paris, France, 28/06/09. https://doi.org/10.1145/1601966.1601981

OcVFDT : One-class very fast decision tree for one-class classification of data streams. / Li, Chen; Zhang, Yang; Li, Xue.

Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09. Association for Computing Machinery (ACM), 2009. p. 79-86 1601981.

Research output: Chapter in Book/Report/Conference proceeding › Conference Paper › Research › peer-review

TY - GEN

T1 - OcVFDT

T2 - 3rd International Workshop on Knowledge Discovery from Sensor Data

AU - Li, Chen

AU - Zhang, Yang

AU - Li, Xue

N1 - Conference code: 3

PY - 2009

Y1 - 2009

N2 - Current research on data stream classification mainly focuses on supervised learning, in which a fully labeled data stream is needed for training. However, fully labeled data streams are expensive to obtain, which make the supervised learning approach difficult to be applied to real-life applications. In this paper, we model applications, such as credit fraud detection and intrusion detection, as a one-class data stream classification problem. The cost of fully labeling the data stream is reduced as users only need to provide some positive samples together with the unlabeled samples to the learner. Based on VFDT and POSC4.5, we propose our OcVFDT (One-class Very Fast Decision Tree) algorithm. Experimental study on both synthetic and real-life datasets shows that the OcVFDT has excellent classification performance. Even 80% of the samples in data stream are unlabeled, the classification performance of OcVFDT is still very close to that of VFDT, which is trained on fully labeled stream.

AB - Current research on data stream classification mainly focuses on supervised learning, in which a fully labeled data stream is needed for training. However, fully labeled data streams are expensive to obtain, which make the supervised learning approach difficult to be applied to real-life applications. In this paper, we model applications, such as credit fraud detection and intrusion detection, as a one-class data stream classification problem. The cost of fully labeling the data stream is reduced as users only need to provide some positive samples together with the unlabeled samples to the learner. Based on VFDT and POSC4.5, we propose our OcVFDT (One-class Very Fast Decision Tree) algorithm. Experimental study on both synthetic and real-life datasets shows that the OcVFDT has excellent classification performance. Even 80% of the samples in data stream are unlabeled, the classification performance of OcVFDT is still very close to that of VFDT, which is trained on fully labeled stream.

UR - http://www.scopus.com/inward/record.url?scp=70450048040&partnerID=8YFLogxK

U2 - 10.1145/1601966.1601981

DO - 10.1145/1601966.1601981

M3 - Conference Paper

AN - SCOPUS:70450048040

SN - 9781605586687

SP - 79

EP - 86

BT - Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09

PB - Association for Computing Machinery (ACM)

Y2 - 28 June 2009 through 28 June 2009

ER -

Li C, Zhang Y, Li X. OcVFDT: One-class very fast decision tree for one-class classification of data streams. In Proceedings of the 3rd International Workshop on Knowledge Discovery from Sensor Data, SensorKDD'09 in Conjunction with the 15th ACM SIGKDD Conference on Knowledge Discovery and Data Mining, KDD-09. Association for Computing Machinery (ACM). 2009. p. 79-86. 1601981 https://doi.org/10.1145/1601966.1601981

OcVFDT: One-class very fast decision tree for one-class classification of data streams

Abstract

Conference

Access to Document

Other files and links

Cite this