S. Dumais. Using svms for text categorization. IEEE Intelligent Systems, 13(4), 1998.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....as Figure 1 illustrates. Fig. 1. Structure of a typical document categorization system Practically all currently popular classification algorithms have been used in document categorization systems. These include Naive Bayes [4] 5] decision trees [6] neural nets, kNN and finally SVM [7] [8] which seem to achieve best performance on standard test corpora. The paper [9] compares the performance of above algorithms applied to text data, concluding that SVM , kNN and Rocchio obtain best results, while worst performance seems to be achieved by Naive Bayes, and Neural Nets. No matter how ....

Dumais S., "Using SVMs for text categorization", IEEE Intelligent Systems, July/August 1998.

....support vector machines (SVMs) and neural networks. In text classification, each document in a set is represented as a vector of words. New documents are assigned to predefined categories using textual content. Recently, SVMs have been shown to yield promising results for text categorization [6, 7, 11]. Although programming languages are written in a manner different from natural languages and have some commented information, programming languages have specific keywords and Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without ....

Dumais, S. T. "Using SVMs for text categorization." IEEE Intelligent Systems Magazine, Trends and Controversies, 13(4):21-23, 1998.

....as a module in a larger system (e.g. 19, 30, 35] The methods used for document categorization can be borrowed from Machine Learning. Popular methods include Naive Bayes (e.g. 17, 33] Decision Trees (e.g. 34] Neural Networks, Nearest Neighbor (e.g. 8] and Support Vector Machines (e.g. [9, 33]) In all these methods, a collection of precategorized documents is used to train a statistical model of word or phrase use and then the statistical model is applied to uncategorized documents. Before the training and the actual categorization, there are two preliminary steps: 1) feature ....

....very rare words that burden the classifier more than that they add discrimination power. See for instance [15] As one example, the Support Vector Machine (SVM) is a relatively new but promising technique for pattern categorization and it has been successfully applied to text (see e.g. [9]) In an SVM , documents are represented as points in a vector space, where the dimensions are the selected features. Based on the training document vectors, the SVM finds the (unique) hyperplane that minimizes the expected generalization error. It does this by maximizing the shortest distance ....

Dumais S: Using SVMs for text categorization. IEEE Intelligent Systems 13(4), 1998 pp 21-23.

....document. In this regard, a number of statistical and machine learning methods have been applied to this domain in recent years including nearest neighbor classification [17] naive Bayes with EM (Expectation Maximization) 12] 15] Winnow with active learning [10] Support Vector Machines (SVMs) [7], 8] Maximum Entropy model [14] It is, however, slightly different with others in that our task deal with more objective and confined categories, such as good or bad for a company s financial outlook than categorization of news articles into politics or economics. The proposed method, ....

S. Dumais. Using svms for text categorization. IEEE Intelligent Systems, 13(4), 1998.

....content of the document. Numerous statistical and machine learning methods have been applied to this domain in recent years including nearest neighbor classi cation [17] naive Bayes with EM (Expectation Maximization) 12] 14] Winnow with active learning [10] Support Vector Machines (SVMs) [6], 8] Maximum Entropy model [13] It is, however, slightly di erent with others in that our task deal with more objective and con ned classes, such as good or bad for a company s nancial outlook than classi cation of news articles into politics or economics. The text classi cation ....

Dumais, S., Using SVMs for text categorization, IEEE Intelligent Systems, Vol. 13, No. 4, 1998.

....takes approximately 2.2 times as long as training a single SVM on an entire training set. All of the results in this chapter are presented using three fold cross validation. Even with cross validated unbiased training data, the sigmoid can still be over t. For example, in the Reuters data set [5, 8], some of the categories have very few positive examples which are linearly separable from all of the negative examples. Fitting a sigmoid for these SVMs with maximum likelihood will simply drive the parameter A to a very large negative number, even if the positive examples are reweighted. There ....

....Linear 9603 3299 0.08 300 118 Adult Linear 32562 16282 0.05 123 1 Adult Quadratic 1605 16282 0.3 123 1 Web Linear 49749 21489 1.0 300 1 Web Quadratic 2477 21489 10.0 300 1 Table 1: Experimental Parameters tasks were used. The rst task is determining the category of a Reuters news article [5, 8]. The second task is the UCI Adult benchmark of estimating the income of a household given census form data [13] where the input vectors are quantized [15] The third task is determining the category of a web page given key words in the page [15] The Reuters task is solved using a linear SVM, ....

S. Dumais. Using SVMs for text categorization. IEEE Intelligent Systems, 13(4), 1998. In: M.A. Hearst, B. Scholkopf, S. Dumais, E. Osuna, and J. Platt: Trends and Controversies | Support Vector Machines.

....on the benchmark Reuters corpus of newswire stories, where the top performing methods include kNN and the Linear Least Squares Fit mapping by Yang et al. 31] Generalized Instance Sets by W. Lam et al. 18] decision trees with boosting by Weiss et al. 27] Support Vector Machines by Joachims[17, 11], and neural networks by Wiener et al. 28] Other methods that performed less well in TC include Naive Bayes classifiers, decision trees without boosting, and rule induction algorithms[31, 32] We chose kNN for event tracking because, in addition to its generally good performance, it makes the ....

S.T. Dumais. Using svms for text categorization. In Support Vector Learning In: IEEE Intelligent Systems, pages July--August, 13(4), 1998.

No context found.

S. Dumais. Using svms for text categorization. IEEE Intelligent Systems, 13(4), 1998.

No context found.

S. Dumais, Using SVMs for text categorization, IEEE Intell. Systems 13(4) (1998) 2123.

No context found.

S. Dumais. Using svms for text categorization. IEEE Intelligent Systems, 13(4), 1998.

No context found.

S. Dumais, Using SVMs for text categorization, IEEE Intelligent Systems 13 (4) (1998) 21 -- 23.

No context found.

Dumais, S.: Using SVMs for text categorization. IEEE Intelligent Systems (1998) 21--23

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