Lacking Labels in the Stream: Classifying Evolving Stream Data with Few Labels. Woolam, C., Masud, M. M., & Khan, L. In Proceedings of the 18th International Symposium on Foundations of Intelligent Systems, of ISMIS '09, pages 552–562, Berlin, Heidelberg, August, 2009. Springer-Verlag. ![Lacking Labels in the Stream: Classifying Evolving Stream Data with Few Labels [link]](https://bibbase.org/img/filetypes/link.svg "link")
Paper doi abstract bibtex This paper outlines a data stream classification technique that addresses the problem of insufficient and biased labeled data. It is practical to assume that only a small fraction of instances in the stream are labeled. A more practical assumption would be that the labeled data may not be independently distributed among all training documents. How can we ensure that a good classification model would be built in these scenarios, considering that the data stream also has evolving nature? In our previous work we applied semi-supervised clustering to build classification models using limited amount of labeled training data. However, it assumed that the data to be labeled should be chosen randomly. In our current work, we relax this assumption, and propose a label propagation framework for data streams that can build good classification models even if the data are not labeled randomly. Comparison with state-of-the-art stream classification techniques on synthetic and benchmark real data proves the effectiveness of our approach.
@inproceedings{woolam_lacking_2009,
address = {Berlin, Heidelberg},
series = {{ISMIS} '09},
title = {Lacking {Labels} in the {Stream}: {Classifying} {Evolving} {Stream} {Data} with {Few} {Labels}},
isbn = {978-3-642-04124-2},
shorttitle = {Lacking {Labels} in the {Stream}},
url = {https://doi.org/10.1007/978-3-642-04125-9_58},
doi = {10.1007/978-3-642-04125-9_58},
abstract = {This paper outlines a data stream classification technique that addresses the problem of insufficient and biased labeled data. It is practical to assume that only a small fraction of instances in the stream are labeled. A more practical assumption would be that the labeled data may not be independently distributed among all training documents. How can we ensure that a good classification model would be built in these scenarios, considering that the data stream also has evolving nature? In our previous work we applied semi-supervised clustering to build classification models using limited amount of labeled training data. However, it assumed that the data to be labeled should be chosen randomly. In our current work, we relax this assumption, and propose a label propagation framework for data streams that can build good classification models even if the data are not labeled randomly. Comparison with state-of-the-art stream classification techniques on synthetic and benchmark real data proves the effectiveness of our approach.},
urldate = {2022-03-28},
booktitle = {Proceedings of the 18th {International} {Symposium} on {Foundations} of {Intelligent} {Systems}},
publisher = {Springer-Verlag},
author = {Woolam, Clay and Masud, Mohammad M. and Khan, Latifur},
month = aug,
year = {2009},
pages = {552--562},
}
Downloads: 0
{"_id":"C9fPoDxNYbwAmbEp3","bibbaseid":"woolam-masud-khan-lackinglabelsinthestreamclassifyingevolvingstreamdatawithfewlabels-2009","author_short":["Woolam, C.","Masud, M. M.","Khan, L."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","address":"Berlin, Heidelberg","series":"ISMIS '09","title":"Lacking Labels in the Stream: Classifying Evolving Stream Data with Few Labels","isbn":"978-3-642-04124-2","shorttitle":"Lacking Labels in the Stream","url":"https://doi.org/10.1007/978-3-642-04125-9_58","doi":"10.1007/978-3-642-04125-9_58","abstract":"This paper outlines a data stream classification technique that addresses the problem of insufficient and biased labeled data. It is practical to assume that only a small fraction of instances in the stream are labeled. A more practical assumption would be that the labeled data may not be independently distributed among all training documents. How can we ensure that a good classification model would be built in these scenarios, considering that the data stream also has evolving nature? In our previous work we applied semi-supervised clustering to build classification models using limited amount of labeled training data. However, it assumed that the data to be labeled should be chosen randomly. In our current work, we relax this assumption, and propose a label propagation framework for data streams that can build good classification models even if the data are not labeled randomly. Comparison with state-of-the-art stream classification techniques on synthetic and benchmark real data proves the effectiveness of our approach.","urldate":"2022-03-28","booktitle":"Proceedings of the 18th International Symposium on Foundations of Intelligent Systems","publisher":"Springer-Verlag","author":[{"propositions":[],"lastnames":["Woolam"],"firstnames":["Clay"],"suffixes":[]},{"propositions":[],"lastnames":["Masud"],"firstnames":["Mohammad","M."],"suffixes":[]},{"propositions":[],"lastnames":["Khan"],"firstnames":["Latifur"],"suffixes":[]}],"month":"August","year":"2009","pages":"552–562","bibtex":"@inproceedings{woolam_lacking_2009,\n\taddress = {Berlin, Heidelberg},\n\tseries = {{ISMIS} '09},\n\ttitle = {Lacking {Labels} in the {Stream}: {Classifying} {Evolving} {Stream} {Data} with {Few} {Labels}},\n\tisbn = {978-3-642-04124-2},\n\tshorttitle = {Lacking {Labels} in the {Stream}},\n\turl = {https://doi.org/10.1007/978-3-642-04125-9_58},\n\tdoi = {10.1007/978-3-642-04125-9_58},\n\tabstract = {This paper outlines a data stream classification technique that addresses the problem of insufficient and biased labeled data. It is practical to assume that only a small fraction of instances in the stream are labeled. A more practical assumption would be that the labeled data may not be independently distributed among all training documents. How can we ensure that a good classification model would be built in these scenarios, considering that the data stream also has evolving nature? In our previous work we applied semi-supervised clustering to build classification models using limited amount of labeled training data. However, it assumed that the data to be labeled should be chosen randomly. In our current work, we relax this assumption, and propose a label propagation framework for data streams that can build good classification models even if the data are not labeled randomly. Comparison with state-of-the-art stream classification techniques on synthetic and benchmark real data proves the effectiveness of our approach.},\n\turldate = {2022-03-28},\n\tbooktitle = {Proceedings of the 18th {International} {Symposium} on {Foundations} of {Intelligent} {Systems}},\n\tpublisher = {Springer-Verlag},\n\tauthor = {Woolam, Clay and Masud, Mohammad M. and Khan, Latifur},\n\tmonth = aug,\n\tyear = {2009},\n\tpages = {552--562},\n}\n\n","author_short":["Woolam, C.","Masud, M. M.","Khan, L."],"key":"woolam_lacking_2009","id":"woolam_lacking_2009","bibbaseid":"woolam-masud-khan-lackinglabelsinthestreamclassifyingevolvingstreamdatawithfewlabels-2009","role":"author","urls":{"Paper":"https://doi.org/10.1007/978-3-642-04125-9_58"},"metadata":{"authorlinks":{}}},"bibtype":"inproceedings","biburl":"https://bibbase.org/zotero/mh_lenguyen","dataSources":["iwKepCrWBps7ojhDx"],"keywords":[],"search_terms":["lacking","labels","stream","classifying","evolving","stream","data","few","labels","woolam","masud","khan"],"title":"Lacking Labels in the Stream: Classifying Evolving Stream Data with Few Labels","year":2009}