An incremental RVR learning method for aviation safety prediction

作者：Ming Chang^*; Xuqun You

来源：International Journal of Digital Content Technology and Its Applications, 2012, 6(15): 381-387.

DOI：10.4156/jdcta.vol6.issue15.43

• Aviation safety
• Incremental learning
• Input vector
• Learning methods
• Prediction model
• Support vector
• Training sample

It is well-known that aviation safety prediction is very important to prevent the aviation accidents. Here, we present an incremental relevance vector regression learning method for aviation safety prediction. The two, three dimensional input vectors of the training samples are employed to construct the training samples respectively in this study. The traditional RVR learning method is used to compare with the incremental RVR learning method. we can see that the prediction effects of the prediction models trained by the training samples with three dimensional input vector are better than those of the prediction models trained by the training samples with two dimensional input vector. In addition, the incremental RVR learning method has the better aviation accidents prediction ability than the traditional RVR learning method.

• 出版日期2012
• 单位陕西师范大学