Abstract：To solve the classification issues of DNA microarray data and enhance the recognition rate, an AdaBoostbased selective ensemble learning method
was proposed with evolvable hardware (EHW) multiple classifiers. At the system ensemble stage, two improved AdaBoost algorithms were introduced. A sample
labeling method was used to improve the effective capacity of sampling, and a selective ensemble strategy was employed to directly promote the classification
precision of combined classifier. The experiments were completed on acute leukemia, lung cancer and colon cancer. The results show that the average
accuracies of the proposed AdaBoostbased EHW for acute leukemia, lung cancer and colon cancer dataset are 97.06％, 99.32％ and 94.44％, respectively. The
proposed scheme achieves higher classification rate and lower hardware cost than the traditional EHW ensemble learning methods.
2017, Vol. 38 
