DOI: 10.5176/978-981-08-9266-1_DAMD13
Authors: Yu-Mei Hung, Wei-Zhan Hung, Ming-Fu Hsu and Ping-Feng Pai
Abstract:
Adaboost is an emerging algorithm used to reduce the error of learning algorithm. However, there are not widely discussed about the application of Adaboost for predicting students’ mathematics and science achievement from the Trends in International Mathematics and Science Study (TIMSS). To prevent cause confusion for the mining procedure, resulting in an unreliable outcome, the raw data undergoes data-preprocessing. ReliefF approach is applied to identify the informative attributes which would enhance the performance in terms of classification capability and reduce the computational complexity. Adaboost algorithm can be viewed as a data post-processing, the inherent mechanism is modifying the weight to attain the superior outcome. Hence, the investigation proposed ReliefFAda model which contained two data processing procedures to predict the student’s mathematics and science achievement. The finding could give directions for researchers, educationists and parents to assess the teaching skills and improve the education environment.
Keywords: Mathematics and science achievement, TIMSS, ReliefF, Feature selection, Adaboost, organization culture
Updating...