P. Cheeseman, J. Kelly, and M. Self. AutoClass: A bayesian classification system. In ICML'88, 1988.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....separate profile for each input class. The partitioning can be done using any attribute of the input that may influence performance, such as size or a complexity measure. Input classes of similar performance can also be derived automatically using Bayesian statistics by programs such as Autoclass [3]. In this paper, we consider conditioning on the input quality. A conditional performance profile therefore consists of a mapping from input quality and run time to probability distribution of output quality: Definition 2.5 The conditional performance profile (CPP) of an algorithm ,A is a ....

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor and D. Freeman, Autoclass: a Bayesian classification system, In Proceedings of the Fifth International Conference on Machine Learning, (1988).

....separate profile for each input class. The partitioning can be done using any attribute of the input that may influence performance, such as size or a complexity measure. Input classes of similar performance can also be derived automatically using Bayesian statistics by programs such as Autoclass [Cheeseman et al. 1988] . Additional types of conditional performance profiles, used for compilation purposes, are discussed in Chapter 5. Definition 4.3 The expected performance profile (EPP) of an algorithm that maps computation time to the expected quality of the results. An expected performance profile is ....

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor and D. Freeman. Autoclass: a Bayesian classification system. In Proceedings of the Fifth International Conference on Machine Learning, 1988.

....based on the similarity of their behavior over different trading days. Note that, when viewing the data from the perspective of learning a Bayesian network or a module network, the instances are trading days and their attributes are stocks. We can use any standard clustering procedure (e.g. [2]) to come up with this initial clustering. We choose to use a procedure that is suitable to our problem, in that it evaluates a partition of variables into modules by measuring the extent to which the module model is a good fit to the data of the variables in the module. This algorithm can be ....

....modules in the network with 50 modules for annotations representing various sectors to which each stock belongs. We found significant enrichment for 21 such annotations, covering a wide variety of sectors. We also compared these results to the clusters of stocks obtained from applying Autoclass [2] to the data. Here, as we described above, each instance corresponds to a stock and is described by 273 random variables, each representing a trading day. In 20 of the 21 cases, the enrichment was far more significant in the modules learned using module networks compared to the one learned by ....

[Article contains additional citation context not shown here]

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. Autoclass: a Bayesian classification system. In ML '88. 1988.

....to certain criterion. A commonly accepted object clustering criterion is the principle of conceptual clustering [18] clustering a set of objects in an attempt to maximize intraclass similarity and inter class differences. Data clustering has been studied in statistics [11] machine learning [13, 18, 19], and databases [42, 15] with different methods and different emphases. Previous approaches, probability based (like most approaches in machine learning) or distance based (like many methods in statistics) do not adequately consider the cases the data sets can be too large to fit into main ....

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. Autoclass: a bayesian classification system. In Proc. Fifth Int. Conf. on Machine Learning, pages 54-64, San Mateo, CaJifornia, 1988.

....teams when playing each other. Then, using C4.5, rules are extracted to predict where a pass will go to. The input conditions are the starting location of the pass and the distance and angle of all players. Note that the starting and ending locations of the passes are discretized using Autoclass C[1]. These rules are then translated into coach language advice in the following fashion. For the team to be imitated (Brainstormers for the coach competition) the rules are translated into pass to region actions with the appropriate conditions. For the opponent (Gemini for the coach ....

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. Autoclass: A bayesian classification system. In ICML-88, pages 54-64, San Francisco, June 1988.

....observes the passes of teams in previous games in order to learn rules which capture some of these passing patterns. These rules can then be used either to imitate a team, or to predict the passes that an opponent will do. The rule learning uses a combination of clustering (using Autoclass C [3]) to create regions on the field and C4.5 [12] to generate rules describing the passing behaviour of a team. The attributes for the rules are the locations of the passer and receiver (using the regions learned from clustering) and the 1 3 5 7 realtive position of all teammates and ....

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. Autoclass: A bayesian classification system. In ICML-88, pages 54-64, San Francisco, June 1988. Morgan Kaufmann.

....assigning the occurrence the sense of that centtold. In summary, the disambiguation algorithm proposed here has three phases. First, a number of context vectors of a target word are computed from a training set. The context vectors are clustered, with each cluster ideally correspond SAutoClass [2] is used for some o the words in section Case Studies . word senses g correct 1 2 3 total capitaIls goods seat of government 5 2 96 92 95 interesi s special attention financial 15 3 94 92 93 raolion s movement proposal 0 2 92 91 92 plani s factory living being 4 13 94 88 92 ruling ....

Peter Cheeseman, James Kelly, Matthew Self, John Stutz, Will Taylor, and Don Freeman. AutoClass: A Bayesian classification system. In Proceedings of the Fifth International Conference on Machine Learning, 1988.

....(level 3) In a model based approach, parametric cluster models are fitted to the obtained features. Here, the data D is given by the set of points in the parameter space of the features of level 1 and 2. The parameter q is given by the parameters of the cluster model. With the AutoClass package [6] a powerful implementation of Bayesian classification is publicly available. Geometric features (level 4) can be obtained by describing the image segments on level 3. However, they can also be extracted directly from the image data on level 0. A sophisticated Bayesian approach can even take into ....

P. Cheeseman et. al. AutoClass: A Bayesian classification system. In Proceedings of the Fifth Int. Conference on Machine Learning, pages 54--64. 1988.

....technique with a partial linear model. The key idea is to obtain partial linear models for clusters of data instead of fitting a single model to all data. The first step of our methodology consists of obtaining a clustering of the data. For this purpose we have used system AUTOCLASS C 1 [5, 6]. The main motivation for this choice was the fact that AUTOCLASS C provides the features that we need to implement our method, namely, cluster membership probabilities and automatic choice of the number of clusters. As the goal of this clustering stage is to find groups of similar training ....

Cheeseman,P., Kelly,J., Self, M., Stutz,J. : Autoclass: a bayesian classification system. In Proceedings of the 5 th International Conference on Machine Learning. Morgan Kaufmann, 1988.

....Gordon, 1993 ] Conditional probability distributions used by Bayesian classifiers are derived from the frequency distributions of attribute values and reflect the 15 Figure 2.1: Meta learning process. likelihood of a certain instance belonging to a particular classification [ Duda Hart, 1973; Cheeseman et al. 1988 ] Implicit decision rules classify according to maximal probabilities. Most of the current research has concentrated on determining the learning algorithm that best fits the target data mining application to compute the best possible classification model. Recently, however, there has been ....

Cheeseman, P.; Kelly, J.; Self, M.; Stutz, J.; Taylor, W.; and Freeman, D. 1988. Autoclass: A bayesian classification system. In Proc. Fifth Intl. Conf. Machine Learning, 54--64. 129

....location and heading change, requiring the use and manipulation of directional data. Probabilistic models are widely used within the AI community. Such models may allow continuous probabilities, as demonstrated in work on Bayesian networks [7] hidden Markov models [5, 8] probabilistic clusters [2] and stochastic maps [19] to name a few. However, the assumption underlying all the above work is that continuous distributions are linear that is distributions that assign density to each point on the real line so that the area under the density curve, integrated over the whole real line, ....

....are based on distributions that are either discrete or continuous along the real line, rather than circular. It is important to be aware of the need for circular distributions as well as of their existence. Moreover, it would be useful to have widely used applications such as Autoclass [2] support such distributions. A problematic aspect of directional data which manifests itself when learning maps and models for robot navigation is that of cumulative rotational errors. In the context of our work we have demonstrated that the use of relative co ordinate systems rather than global ....

P. Cheeseman, et al. Autoclass: A Bayesian classification system. In J. W. Shavlik, T. G. Dietterich, eds., Readings in Machine Learning. Morgan-Kaufmann, 1990.

No context found.

P. Cheeseman, J. Kelly, and M. Self. AutoClass: A bayesian classification system. In ICML'88, 1988.

No context found.

Cheeseman, P., Kelly, J., Self, M., Stutz, J., Taylor, W., & Freeman, D. (1988). Autoclass: A Bayesian classification system. In Proceedings of the Fifth International Conference on Machine Learning, Ann Arbor, MI.

No context found.

Cheeseman, P.; Kelly, J.; Self, M.; Stutz, J.; Taylor, W.; and Freeman, D. 1988. Autoclass: A Bayesian classification system. In Proceedings of the Fifth International Workshop on Machine Learning.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman, `Autoclass: A bayesian classification system', in Proceedings of the 5th Internatinal Conference on Machine Learning, Ann Arbor, MI, pp. 54--56, (1988).

No context found.

P. Cheeseman, J. Stutz, J. Taylor, M. Self, and J. Kelley. Autoclass: A bayesian classification system. Proceedings of the 5th International Conference on Machine Learning, 1988.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman, "AutoClass: A Bayesian Classification System," in Readings in Machine Learning, The Morgan Kaufman Series in Machine Learning, J. W. Shavlik and T. G. Dietterich, Eds. San Mateo, California: Mogan Kaufmann Publishers, Inc., pp. 296-306.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. AutoClass: A bayesian classification system. In Fifth International Conference on Machine Learning, pages 54--64, Ann Arbor, MI, 1988.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman, "AutoClass: A Bayesian Classification System," in Readings in Machine Learning, The Morgan Kaufman Series in Machine Learning, J. W. Shavlik and T. G. Dietterich, Eds. San Mateo, California: Mogan Kaufmann Publishers, Inc., pp. 296-306.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W.Taylor, and D. Freeman, Autoclass: A bayesian classification system, Proceedings of the Fifth International Conference on Machine Learning (Ann Arbor, MI.), Morgan Kaufmann Publishers, 1988, pp. 54--64.

No context found.

P. Cheeseman, J. Kelly, M. Self, "AutoClass: a Bayesian classification system", Proc. of 5th Int'l Conf. on Machine Learning, Morgan Kaufman, June 1988.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor and D. Freeman. "AUTOCLASS: a Bayesian Classification System". Procc. of 5th Int. Conference on Machine Learning (ICML- 1988), pp. 54-64. Ann Arbor, MI, 1988.

No context found.

P. Cheeseman, J. Kelly, M. Self, J. Stutz, W. Taylor, and D. Freeman. Autoclass: A bayesian classification system. In Proceedings of the 5 t h International Workshop on Machine Learning, pages 54--64, 1988.

No context found.

Cheeseman P., Kelly J., Self M., Stutz J., Taylor W., Freeman D.: AutoClass: A Bayesian classification system, In Proceedings of the Fifth International Conference on Machine Learning (ML'88), 54-64, 1988.

No context found.

P. Cheeseman et al., "AUTOCLASS: A Bayesian Classification System," Proc. ICML-86, Morgan Kaufmann, San Francisco, 1988, pp. 54--64.

First 50 documents  Next 50

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