D. T. Davis and J. N. Hwang, "Attentional focus training by boundary region data selection," in Proc. Int. Joint Conf. Neural Networks, (IJCNN'92), Baltimore, MD, June 1992, pp. 676--681.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....in the greatest potential improvement in classification accuracy. Boundary marking is a technique whereby evenly spaced boundary points can be generated for eventual presentation to the oracle. Query based learning has been applied to a number of applications, include cytology screening [16] [17], power system security assessment [18] 20] and classification of incomplete data [23] Example 2 Query Based Learning Applied to Power System Security Assessment: An application of neural network query learning using boundary marking is in the field of power system security assessment. This ....

D. T. Davis and J. N. Hwang, "Attentional focus training by boundary region data selection," in Proc. Int. Joint Conf. Neural Networks, (IJCNN'92), Baltimore, MD, June 1992, pp. 676--681.

....in the greatest potential improvement in classification accuracy. Boundary marking is a technique whereby evenly spaced boundary points can be generated for eventual presentation to the oracle. Query based learning has been applied to a number of applications, include cytology screening [16] [17], power system security assessment [18] 19] 20] and classification of incomplete data [23] Example 2: Query based Learning Applied to Power System Security Assessment An application of neural network query learning using boundary marking is in the field of power system security assessment. ....

D. T. Davis and J. N. Hwang, "Attentional Focus Training by Boundary Region Data Selection", International Joint Conference on Neural Networks, (IJCNN'92), Baltimore, MD, pp. 676-681, June, 1992.

....whose class membership is uncertain (given the current classifier) are proposed to a user for membership judgment. Most Sampling techniques in machine learning aim at reducing the size of the training set [19] and are not motivated by improving the classification accuracy (except for a few like [10]) Our zoning techniques, on the other hand, are specifically aimed at improving the routing effectiveness and are not motivated by the size of the training corpus. Recently Kwok and Grunfeld have used a sampling technique based on genetic algorithms that selects the best training subset of the ....

D. T. Davis and J.-N. Hwang. Attentional focus training by boundary region data selection. In International Joint Conference on Neural Networks, pages I--676 to I--681, Baltimore, MD, June 7--11 1992.

....of the set of all classifiers consistent with the labeled data: the version space [24] The degree to which this is a problem in practice has not been established. Single classifier approaches have successfully been used in generating arbitrary queries [16] and in sampling from labeled data [8, 25]. Uncertainty sampling with a single classifier can also be viewed as a variation on the heuristic of training on misclassified instances [15, 33, 35] A familiar example of this is windowing, which appeared in Quinlan s first paper on ID3 [26] was questioned in [36] and re examined in Chapter 6 ....

Daniel T. Davis and Jenq-Neng Hwang. Attentional focus training by boundary region data selection. In International Joint Conference on Neural Networks, pages I--676 to I--681, Baltimore, MD, June 7--11 1992.

....(Angluin 1988) the learner is allowed to construct artificial examples, while selective sampling deals with selection of informative examples from a large set of unclassified data. Selective sampling methods have been developed for various classification learning algorithms: for neural networks (Davis Hwang 1992; Cohn, Atlas, Lander 1994) for the C4:5 rule induction algorithm (Lewis Catlett 1994) and for HMM (Dagan Engelson 1995) The goal of the research described in this paper is to develop a selective sampling methodology for nearest neighbor classification learning algorithms. The Copyright c ....

Davis, D. T., and Hwang, J.-N. 1992. Attentional focus training by boundary region data selection. In IJCNN, volume 1, 676--81. IEEE.

.... approach to uncertainty sampling has several theoretical failings, including underestimation of true uncertainty, and biases caused by nonrepresentative classifiers [9, 10] On the other hand, experiments using a single classifier to make arbitrary queries [14] or select subsets of labeled data [8, 15] have shown substantial speedups in learning. Relevance sampling, which has proven quite effective for text retrieval, also uses a single classifier. 3 An Uncertainty Sampling Algorithm Figure 3 presents an algorithm for uncertainty sampling from a finite set of examples using a single ....

D. T. Davis and J. Hwang. Attentional focus training by boundary region data selection. In International Joint Conference on Neural Networks, pages I--676 to I--681, Baltimore, MD, June 7--11 1992.

....the one shown in figure 1b. Figures 4 and 5 are examples. However visualization is very difficult for high dimensional feature spaces. Many classifiers have been developed using this approach, for example, appendicitis diagnosis[4] detection of bacterial growths[1] and cytological classification[3]. A major goal in developing a classifier for any domain is to achieve the highest possible accuracy in the deployed classifier. Of course one cannot know in advance what the accuracy of the deployed classifier (the true error rate) will be, the best one can do is to estimate this error rate from ....

....trained classifier in the hope that they will be close to the true boundaries. The generation problem in the work reported here has been solved by using a genetic algorithm[5] The manual classification problem has been by passed by choosing problems in which this can be done automatically. In [3] a network inversion search was used to find boundary regions for a cytology classification problem. A 6 increase in classification accuracy is reported using a particular strategy, but there is no indication of how much better or worse this is than alternative strategies. Eberhart [4] suggests ....

[Article contains additional citation context not shown here]

Daniet T. Davis and Jenq-Neng Hwang. Attentional focus training by boundary region data selection. In Proceedings of the International Joint Conference on Neural Networks, pages I--676--I--681, 1992.

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