Research on the Semantic Measurement in Co-word Analysis

Aiming at problems of the "same amount with different qualities'' phenomenon and the lack of semantics in co-occurring terms, this paper proposed a new semantic measurement method in co-word analysis. The method firstly gave different weights to document units based on the Pointwise Mutual Information (PMI) method, and then extended them to the generation process of the Latent Dirichlet Allocation (LDA) model to extract core keywords. Then the word-2vec model was used to transform the Top-N keywords into lowdimensional value distributions, and the sematic correlation among keywords were calculated based on the length of windows. Finally, data from the domain of "deep learning" was used to verify the scientificity and effectiveness of the method. Comparing the results of general co-word analysis with our proposed method in terms of clustering analysis, network parameters, distribution structures and other aspects, we can find that our method is scientific and effective in considering different feasibilities of terms and their semantic correlations.