Penerapan Teknik Data Mining Dengan Metode Smooth Support Vector Machine (SSVM) Untuk Memprediksi Mahasiswa Yang Berpeluang Drop Out (Studi Kasus Mahasiswa Politeknik Negeri Medan)
| Main Author: | Safitri, Habibi Ramdani |
|---|---|
| Other Authors: | Zarlis, Muhammad |
| Format: | Masters |
| Bahasa: | ind |
| Subjects: | |
| Online Access: |
http://repository.usu.ac.id/handle/123456789/31270 |
Daftar Isi:
- Support Vector Machines (SVM) is a new algorithm of data mining techniques, the increasing popularity in machine learning and statistics communities. SVM has been introduced by Vapnik to solve the problem of pattern recognition and nonlinear function estimation. SVM has become the tool of choice for the basic classification problem machine learning and data mining. Unlike traditional methods that minimize the empirical training error, SVM aims at minimizing the upper bound of generalization error through maximizing the margin between the hyperplane separating the data. This can be regarded as the implementation of the principle minimisasi risikostruktur estimates, smoothing method, widely used to solve mathematical programming problems and important applications, which are applied here to generate and solve an infinite reformulation of support vector machines for pattern classification. Although many variants of SVM have been proposed, is still an active research problem in order to improve for a more effective classification. SSVM is a development of the SVM that uses a smoothing technique. This method was first introduced by Leepad atahun 2001. The basic idea is to convert from SSVM SVM primal formulation for non-smooth minimization problem without constraint. Research Support Vector Machine (SSVM) is active in the field of data mining. The author developed a method to improve the accuracy of the results from the database drop-out problem Polytechnic students majoring in particular field of Mechanical Engineering and Energy Conversion Techniques.
- 097038022