CNN-based Evaluation Method of Academic Innovation Effect of American Research Universities

作者：Lina Sun, Sitan Li · 年份：2021 · DOI：10.1109/iaai54625.2021.9699893 · 研究领域：AI and Multimedia in Education、Advanced Technologies in Various Fields、AI and Big Data Applications

Aiming at the problem of fuzzy characteristics of academic achievements in the existing evaluation methods of academic innovation effects of American research universities, resulting in low evaluation accuracy, a CNN-based evaluation method of academic innovation effects of American research universities is designed. First, we obtain innovation influence indicators, calculate the breadth of innovation absorption in papers, and identify the characteristics of academic achievements of American research universities; then, we convert scientific and technological achievements into patent licenses, measure keyword coverage, and finally use CNN to design innovation effect evaluation algorithms. We completed the design of the evaluation method for the academic innovation effect of American research universities. The experimental results show that the average evaluation accuracy of the evaluation method of the academic innovation effect of the American research university and the other two evaluation methods are 73.59%, 62.51%, and 60.28%, respectively, which proves that the American research university integrates CNN technology. The evaluation method of the academic innovation effect is more practical.