A keywords extraction algorithm optimized by integrating co-word analysis

The paper proposed a keywords extraction optimization algorithm based on co-word analysis, named CATKE, to address the problems of the single dimension of word importance evaluation and relatively low precision of extracting keywords in the classic TF-IDF keywords extraction algorithm. The paper first conducted research on co-word analysis and its applications, proposed the formula co-weight(vi) for calculating co-occurrence weight, then proposed the formula key-weight(vi) for calculating key degree by combining the TF-IDF weight and co-occurrence weight of word, and finally proposed a keywords extraction optimization algorithm based on co-word analysis. To evaluate the performance of the proposed algorithm, we conducted comparative experiments with five other baseline methods based on different principles on the same dataset, which included titles and abstracts data from 6400 Chinese scientific papers. The experiments reported the macro-average precision, macro-average recall, macro-average F1-score, ranking quality indicator MAP, and running efficiencies of all the methods to provide a comprehensive evaluation. The experimental results demonstrated the effectiveness of CATKE.