A study of thresholding strategies for text categorization

By Yiming Yang. Annual International ACM SIGIR Conference on Research and Development in Information Retrieval, 2001

416 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Computer Vision and Pattern Recognition, Information Systems

View Paper: https://www.semanticscholar.org/paper/01f3a2d40a748fdd571f9cdb6fc402118919307e

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Computer Vision and Pattern Recognition

In Information Systems

Citations in the corpus

An empirical study of sentiment analysis for chinese documents (2008, background): “It is worth noticing that we do not introduce any thresholds investigated by Yang (2001) because the adjusting of thresholds may incur significant computational costs.”
Optimizing F-measure: A Tale of Two Approaches (2012, methodology): “Though simple, this method has been found to be eﬀective and is commonly applied, for example, in text categorization (Yang, 2001).”
The Enron Corpus: A New Dataset for Email Classi(cid:12)cation Research (2004, methodology): “Therefore, we obtained thresholds using score-based local optimization, known as SCut [12], and evaluated using F1 scores, which measure both precision and recall.”
Hierarchical Bayesian Inference and Recursive Regularization for Large-Scale Classification (2015, methodology): “This provided a better performance than using the usual cut-off of zero and other threshold methods like rcut or scut [Yang 2001].”
Mining Text Data (2012, background): “It has also been suggested in [136] that these thresholding strategies can be used to understand the thresholding strategies of other classifiers which use ranking classifiers.”
RCV1: A New Benchmark Collection for Text Categorization Research (2004, methodology): “To use these models for classification, we use the SCut strategy (Yang, 2001), i.”
Multi-label classification with a reject option (2013, background): “, [16, 17, 18]).”
Classifier chains for multi-label classification (2009, background): “In the context of text categorisation, [37] uses a threshold vector, i.”
An analysis of the relative hardness of Reuters-21578 subsets (2005, methodology): “a validation set by the proportional thresholding method [17, 32], according to which the threshold τi is set to the value such that the proportion of validation examples that are classified into ci is as close as possible to the proportion of training examples that are classified into ci.”
Wikipedia-based Semantic Interpretation for Natural Language Processing (2009, methodology): “• 90 categories with at least one document in the training set and one in the testing set (Yang, 2001).”