Y. Yang. An evaluation of statistical approaches to MEDLINE indexing. In J. J. Cimino, editor, Proceedings of AMIA-96, Fall Symposium of the American Medical Informatics Association, pages 358--362, Washington, US, 1996. Hanley and Belfus. 159  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... the assignment of documents to a predefined set of categories is a very interesting task for Information Access. Nowadays, it is possible to build automatic systems able to classify web pages in Internet directories like Yahoo [14] to index records with respect to library subject headings [27], or to filter spam messages [6] among other applications. In ATC, there exists an interesting opportunity of taking advantage of several knowledge sources. ATC has been the focus of much research in recent years [17, 21, 28] The prevailing tendency in ATC is to apply Machine Learning and ....

Yiming Yang. An evaluation of statistical approaches to MEDLINE indexing. In J. J. Cimino, editor, Proceedings of AMIA-96, Fall Symposium of the American Medical Informatics Association, pages 358--362, Washington, US, 1996. Hanley and Belfus.

.... us to compare results with other published research with the same collection [19] The 119 Heart Diseases categories form a 5 level tree where the rst level corresponds to the root node and the fth level has only leaf We label this set HD 119 in order to stay consistent with the labeling in [42] even though there are actually only 103 categories with positive examples in the training set. 17 Diseases Heart Deseases Heart Defects, Congenital Arrhythmia Endocarditis Tachycardia Heart Septal Defects Ischemia Heart Valve Figure 7. Tree for the 119 categories of the heart diseases ....

....van Rijsbergen s F measure is the best suited measure, but still has the drawback that it might be dicult for the user to de ne the relative importance of recall and precision. We report F 1 values because it allows us to compare results with other researchers who have used the same dataset [16, 19, 42]. In general the F 1 performance is reported as an average value. There are two ways of computing this average: macro average, and micro average. With macro average the F 1 value is computed for each category and these are averaged to get the nal macro averaged F 1 . With micro average we rst ....

[Article contains additional citation context not shown here]

Yang Y. An evaluation of statistical approaches to MEDLINE indexing. In Proceedings of the American Medical Informatic Association (AMIA), pp. 358{ 362, 1996.

....has been studied extensively. Yang (1997) compares 14 categorization algorithms applied to this Reuters corpus as a nonhierarchical categorization problem. Others treating the categories as a hierarchy (Chakrabarti, et al. 1997; Koller Sahami, 1997; Ng et al. 1997; Yang 1996) have also studied this same corpus. Yang (1996, 1997) examines the OHSUMED corpus of medical abstracts. Still others examine the categories as a hierarchy for other corpora, namely the Yahoo Web hierarchy (McCallum et al. 1998; Mladenic Grobelnik, 1998) Precision, recall, and F measure are ....

....Yang (1997) compares 14 categorization algorithms applied to this Reuters corpus as a nonhierarchical categorization problem. Others treating the categories as a hierarchy (Chakrabarti, et al. 1997; Koller Sahami, 1997; Ng et al. 1997; Yang 1996) have also studied this same corpus. Yang (1996, 1997) examines the OHSUMED corpus of medical abstracts. Still others examine the categories as a hierarchy for other corpora, namely the Yahoo Web hierarchy (McCallum et al. 1998; Mladenic Grobelnik, 1998) Precision, recall, and F measure are used by most authors as measures of the ....

[Article contains additional citation context not shown here]

Yang Y. (1996). An Evaluation of Statistical Approaches to MEDLINE Indexing. Proceedings of the AMIA , pp. 358-362.

....cutoff point, which gives the best result, is found for the top ranking categories. ExpNet is evaluated on MEDLINE data set and tested with Linear Least Squares Fit (LSSF) mapping method for comparison. ExpNet showed a performance in Recall and Precision comparable to LSSF. In another work by Yang [2], statistical approaches to text categorization were compared. Eleven methods (k NN, LLSF, WH [10] RIPPER [11] EG [10] NNets [15] SWAP 1 [12] CHARADE [14] WORD (word matching) Rocchio [10] NaiveBayes [13] were analyzed and k NN was chosen as the performance baseline for several ....

....the training instances. 5 Application of k NNFP to Text Categorization As stated in Section 2, in [1] it is reported that k NN method shows a performance in Recall and Precision comparable to LSSF mapping method, and significantly better then other methods tested. Also in a recent work of Yang [2], results of a comparative study for statistical approaches to text categorization shows that k NN method is one of the top performing classifiers and it is the only method that has scaled to the full domain of MEDLINE categories. According to the experiments reported in [3] k NNFP method ....

Y. Yang. An Evaluation of Statistical Approaches to MEDLINE Indexing. In Proceedings of the 1996 Annual Full Symposium of the American Medical Informatics Association (1996 AMIA), 358-362, 1996.

....documents were manually indexed using subject categories (Medical Subject Headings, or MeSH; about 18,000 categories defined) in the National Library of Medicine. The OHSUMED collection has been used with the full range of categories (14,321 MeSH categories actually occurred) in some experiments[17], or with a subset of categories in the heart disease sub domain (HD, 119 categories) in other experiments[8] 3.2 Different versions Table 1 lists the different versions or subsets of Reuters and OHSUMED. Each is referred as a set or collection , and labelled for reference. To examine the ....

Y. Yang. An evaluation of a statistical approaches to medline indexing. In Proceedings of the 1996 Annual Full Symposium of the American Medical Informatics Association (1996 AMIA), pages 358--362, 1996.

No context found.

Y. Yang. An evaluation of statistical approaches to MEDLINE indexing. In J. J. Cimino, editor, Proceedings of AMIA-96, Fall Symposium of the American Medical Informatics Association, pages 358--362, Washington, US, 1996. Hanley and Belfus. 159

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC