Under the background of educational informatization, the widespread adoption of virtual teaching platforms has led to an explosive growth in educational resources. However, the limited study time of vocational college students has increasingly highlighted the contradiction between the two. To address this issue, this research proposes a feature weight evaluation method incorporating the Genetic Algorithm (GA), based on statistical models and semantic analysis, tailored to the characteristics of vocational English test texts. This method introduces four feature factors—word frequency, word length, word co-occurrence, and position—and employs feature weight adjustment coefficients along with the GA to adaptively train and optimize each factor. Building upon this, a multi-threaded concurrent computation approach is adopted to achieve rapid feature factor weight calculation, ultimately constructing a complete G-CIWE keyword extraction algorithm. Experimental results demonstrate that the G-CIWE algorithm outperforms comparative algorithms in accuracy, recall, and F1 score across various text types, reaching 96.71%, 87.63%, and 95.57%, respectively. To sum up, the G-CIWE algorithm proposed in the study exhibits excellent performance and versatility and can be well applied to the English test questions in higher vocational colleges.
Liu et al. (Fri,) studied this question.