A dynamical learning method with SVM and its application on bank slip recognition
-
摘要: 介紹了用支持向量機(SVM)進行動態學習訓練的方法．解決了在機器學習過程中,訓練樣本獲取比較困難,樣本可隨外界條件改變而變化的問題．實踐證明,使用該方法可以動態跟蹤樣本的變化,保證SVM分類器的最優性能．利用該方法設計的銀行票據OCR系統的實際應用說明了該方法的有效性．Abstract: This paper introduces a dynamical learning method using support vector machine (SVM). This method can solve such machine learning problems as the difficulties in gathering training samples and the change of samples with outer environment. It is proved that SVM classifiers can achieve optimal performance after using this method in tracking the change of samples. A bank slip OCR system designed by this method proves the validity.
-
點擊查看大圖
計量
- 文章訪問數: 149
- HTML全文瀏覽量: 37
- PDF下載量: 7
- 被引次數: 0
下載:
