Zhentao Liang

Selected publications

1. JASIST

   Bias against scientific novelty: A pre-publication perspective

   Zhentao Liang, Jin Mao*, and Gang Li*

   Journal of the Association for Information Science and Technology 2023

   Abstract HTML

   Novel ideas often experience resistance from incumbent forces. While evidence of the bias against novelty has been widely identified in science, there is still a lack of large-scale quantitative work to study this problem occurring in the pre-publication process of manuscripts. This paper examines the association between manuscript novelty and handling time of publication based on 778,345 articles in 1,159 journals indexed by PubMed. Measuring the novelty as the extent to which manuscripts disrupt existing knowledge, we found systematic evidence that higher novelty is associated with longer handling time. Matching and fixed-effect models were adopted to confirm the statistical significance of this pattern. Moreover, submissions from prestigious authors and institutions have the advantage of shorter handling time, but this advantage is diminishing as manuscript novelty increases. In addition, we found longer handling time is negatively related to the impact of manuscripts, while the relationships between novelty and 3- and 5-year citations are U-shape. This study expands the existing knowledge of the novelty bias by examining its existence in the pre-publication process of manuscripts.

2. IPM

   Combining deep neural network and bibliometric indicator for emerging research topic prediction

   Zhentao Liang, Jin Mao*, Kun Lu, Zhichao Ba, and Gang Li

   Information Processing & Management 2021

   Abstract HTML

   Predicting emerging research topics is important to researchers and policymakers. In this study, we propose a two-step solution to the problem of emerging topic prediction. The first step forecasts the future popularity score, a novel indicator reflecting the impact and growth, of candidate topics in a time-series manner. The second step selects novel topics from the candidates predicted to be popular in the first step. Terms with domain characteristics are used as candidate topics. Deep neural networks, specifically LSTM and NNAR, are applied with nine features of topics to predict popularity score. We evaluated the models and five baselines on two datasets from two perspectives, i.e., the ability to (1) predict the correct indicator value and (2) reconstruct the optimal ranking order. Two types of training strategies were compared, including a global strategy that trains a model with all topics and two local strategies that train separate models with different groups of topics. Our results show that LSTM and NNAR outperform other models in predicting the value of popularity score measured by MAE and RMSE, while LightGBM is a competitive baseline in ranking the topics in terms of NDCG@20. The performance difference of global and local strategies is not significant. Emerging topics predicted by our approach are compared with those by other methods. A qualitative assessment on nominated emerging topics suggests topics nominated by machine learning methods are more alike than those by the rulebased model. Some important topics are nominated according to a preliminary literature analysis. This study exploited the strengths of both machine learning and bibliometric indicator approaches for emerging topic prediction. Deep neural networks are applied where objective optimization target can be defined and measured. Bibliometric indicator offers an efficient way to select novel topics from candidates. The hybrid approach shows promise in considering various characteristics of emerging topics when making predictions.

3. SCIM

   Finding citations for PubMed: a large-scale comparison between five freely available bibliographic data sources

   Zhentao Liang, Jin Mao*, Kun Lu, and Gang Li

   Scientometrics 2021

   Abstract HTML

   As an important biomedical database, PubMed provides users with free access to abstracts of its documents. However, citations between these documents need to be collected from external data sources. Although previous studies have investigated the coverage of various data sources, the quality of citations is underexplored. In response, this study compares the coverage and citation quality of five freely available data sources on 30 million PubMed documents, including OpenCitations Index of CrossRef open DOI-to-DOI citations (COCI), Dimensions, Microsoft Academic Graph (MAG), National Institutes of Health’s Open Citation Collection (NIH-OCC), and Semantic Scholar Open Research Corpus (S2ORC). Three gold standards and five metrics are introduced to evaluate the correctness and completeness of citations. Our results indicate that Dimensions is the most comprehensive data source that provides references for 62.4% of PubMed documents, outperforming the official NIH-OCC dataset (56.7%). Over 90% of citation links in other data sources can also be found in Dimensions. The coverage of MAG, COCI, and S2ORC is 59.6%, 34.7%, and 23.5%, respectively. Regarding the citation quality, Dimensions and NIH-OCC achieve the best overall results. Almost all data sources have a precision higher than 90%, but their recall is much lower. All databases have better performances on recent publications than earlier ones. Meanwhile, the gaps between different data sources have diminished for the documents published in recent years. This study provides evidence for researchers to choose suitable PubMed citation sources, which is also helpful for evaluating the citation quality of free bibliographic databases.

4. JOI

   Quantifying cross-disciplinary knowledge flow from the perspective of content: Introducing an approach based on knowledge memes

   Jin Mao, Zhentao Liang*, Yujie Cao, and Gang Li

   Journal of Informetrics 2020

   Abstract HTML

   Knowledge flow between disciplines is typically measured through citations among publications. In this study, we quantify cross-disciplinary knowledge diffusion from the novel perspective of content by introducing knowledge memes, a special type of knowledge unit. Diffusion cascade is proposed to model the diffusion process of knowledge memes. By taking Medical Informatics (MI) as an exemplary interdisciplinary discipline, we measure the knowledge relationships between it and four related disciplines. The diffusion patterns of cross-disciplinary memes are also identified by analyzing the network structure of the diffusion cascade. The results present the knowledge relationships among disciplines measured by knowledge memes, which are different from those measured by citations. It is shown that preferential attachment takes effect in cross-disciplinary knowledge meme diffusion. In addition, cross-disciplinary knowledge memes generally originate earlier and have higher impact than the memes of MI. This study provides insights into new approaches to quantifying knowledge relationships among disciplines and furthers the understanding of content diffusion mechanisms through measurable knowledge units.