Jasa bantu mengerjakan tugas kuliah teknik informatika, hubungi Meruya Statistik 0812 1944 9060

Jasa bantu mengerjakan tugas kuliah teknik informatika, hubungi Meruya Statistik 0812 1944 9060

Jasa konsultasi tesis, disertasi S3 bidang informatika, teknik elektro, teknik sipil, dan berbagai macam riset pascasarjana bidang ilmu teknik. 

Informasi lebih lanjut hubungi Meruya Statistik 0812 1944 9060

Resume untuk buku berjudul : “Data Mining: Theory, Methodology, Techniques, and Applications”

Graham J. Williams Simeon J. Simoff (Eds.)

Resume untuk artikel berjudul : Generality Is Predictive of Prediction Accuracy

Geoffrey I. Webb1 and Damien Brain2

1 Faculty of Information Technology,

Monash University, Clayton, Vic 3800, Australia

webb@infotech.monash.edu.au

2 UTelco Systems,

Level 50/120 Collins St Melbourne, Vic 3001, Australia

In  knowledge acquisition, the  classification    rule   can  achieve high accuracy on the training data.  However, there is another trade-off that will also be inherent  since the more specific rule will make fewer predictions on unseen cases. It is also known that a  classifier has an option of not making predictions   to create a system that makes fewer decisions of higher expected quality.  When the accuracy of the rules on the training data is high, specializing the rules can be expected to raise their accuracy on unseen data towards that obtained on the training data.

Where a classifier must always make decisions and maximization of prediction accuracy is desired, the rules for the class that occurs most frequently should be generalized at the expense of rules for alternative classes. This is because as each rule is generalized it will trend towards the accuracy of a default rule for that class, which will be highest for rules of the most frequently occurring class.

It is also the alternative sources of information that might be brought to bear upon such decisions. We have emphasized that our hypotheses relate only to contexts   to distinguish between the expected accuracy of two rules other than their relative generality  to derive such evidence from training data. 

During knowledge acquisition,  multiple alternative potential rules all appear equally credible.  In   comparison to the   general rule, the accuracy of the   unseen cases will tend to be closer to the accuracy obtained on training data. 

  the accuracy of   classification rules  is  likely to be closer to the accuracy on unseen data of a default rule for the class than will the accuracy on unseen data of the more specific rule. By using  classification rules formed by C4.5rules and random classification rules may develop    learning biases based on rule generality that do not rely upon prior domain knowledge, and may be sensitive to alternative knowledge acquisition objectives, such as trading-off accuracy for cover.  The  frequent existence of rule variants between which traditional rule quality metrics, such as an information measures, could not be distinguished. 

In   knowledge acquisition , the  classification rules  sometimes uses the training data.   If we are selecting rules to use for some decision making task, we must select between such rules with identical performance on the training data. To do so, it needs learning algorithms   that learn rule sets for the purpose of prediction   by making arbitrary choices between rules with equivalent performance on the training data. This masking   in machine learning  provides support for identification and selection between such rule variants.

References

1.            Mitchell, T.M.: Version spaces: A candidate elimination approach to rule learning. In: Proceedings of the Fifth International Joint Conference on Artificial Intelligence.(1977) 305–310

2.            Mitchell, T.M.: The need for biases in learning generalizations. Technical Report CBM-TR-117, Rutgers University, Department of Computer Science, New Brunswick, NJ (1980)

3.            Webb, G.I.: Integrating machine learning with knowledge acquisition through direct interaction with domain experts. Knowledge-Based Systems 9 (1996) 253–266

4.            Webb, G.I., Wells, J., Zheng, Z.: An experimental evaluation of integrating machine learning with knowledge acquisition. Machine Learning 35 (1999) 5–24

5.            Wolpert, D.H.: On the connection between in-sample testing and generalization error. Complex Systems 6 (1992) 47–94

6.            Schaffer, C.: A conservation law for generalization performance. In: Proceedings of the 1994 International Conference on Machine Learning, Morgan Kaufmann (1994)

7.            Rendell, L., Seshu, R.: Learning hard concepts through constructive induction: Framework and rationale. Computational Intelligence 6 (1990) 247–270

8.            Webb, G.I.: Further experimental evidence against the utility of Occam’s razor. Journal of Artificial Intelligence Research 4 (1996) 397–417

9.            Quinlan, J.R.: C4.5: Programs for Machine Learning. Morgan Kaufmann, San Mateo, CA (1993)

10.          Webb, G.I.: OPUS: An efficient admissible algorithm for unordered search. Journal of Artificial Intelligence Research 3 (1995) 431–465

11.          Blake, C., Merz, C.J.: UCI repository of machine learning databases. [Machine-readable data repository]. University of California, Department of Information and Computer Science, Irvine, CA. (2004)

12.          Pazzani, M.J., Murphy, P., Ali, K., Schulenburg, D.: Trading off coverage for accuracy in forecasts: Applications to clinical data analysis. In: Proceedings of the AAAI Symposium on Artificial Intelligence in Medicine. (1994) 106–110

13.          Compton, P., Edwards, G., Srinivasan, A., Malor, R., Preston, P., Kang, B., Lazarus, L.: Ripple down rules: Turning knowledge acquisition into knowledge maintenance. Artificial Intelligence in Medicine 4 (1992) 47–59

14.          Blumer, A., Ehrenfeucht, A., Haussler, D., Warmuth, M.K.: Occam’s Razor. Information Processing Letters 24 (1987) 377–380

15.          Domingos, P.: The role of Occam’s razor in knowledge discovery. Data Mining and Knowledge Discovery 3 (1999) 409–425