Evaluating book impacts via integrating multi-source reviews

Citations and alternative metrics are commonly used to assess impacts of books. However, such metrics often cannot reflect detailed evaluation information, as they are often frequency-level metrics and neglect content information about books. Existing research reveals that book reviews are useful sources for measuring book impacts, which can provide content-level supplementary information for evaluation. Meanwhile, book reviews are generated from multiple sources, including academic reviews from experts and online reviews from the public. These multi-source reviews reveal opinions on books from different dimensions. Thus, evaluating book impact based on single category of reviews are difficult to get fully assessment results. Hence, this paper assesses book impacts by integrating academic and online reviews, so as to measuring book impacts comprehensively. Specifically, we firstly collected books’ academic reviews and online reviews. Then, we extracted multi-dimensional metrics via finer-grained mining on book reviews. Finally, we compared assessment results based on book review integration and library information to verify assessment results. Experimental results infer that book review integration can complement evaluation information for book impact assessment, and thus improve the book evaluation system and methods.