In this paper, we introduce a new approach based on properties of graph grammars to detect conflicts of interest (COIs) in a field represented in the form of a social network. The approach consists of specializing the adaptive star graph grammar (ASGG) of Drewes et al. (Theor Comput Sci 411:3090–3109, 2010) to express kind of subgraphs that we call \(K_4\)-type tournament graphs, corresponding to COIs, that cannot be generated by the node replacement graph grammar. This approach, called graph grammar and \(K_{4}\)-type tournament-based approach to detect conflicts of interest \((GGK_{4}T-COIs)\), is applied to detect COIs in the review process of papers accepted in an international conference which is represented through a social network. In this contribution, the principle of the used graph grammar is not to consider all the generated language but only subgraphs with some properties (corresponding to special graph queries), which identify parts of the social network representing COIs. For evaluating the performances and the efficiency of our proposition, experimentations have been done by comparing it with concurrent methods in the literature. The obtained results have shown that the approach GG\(K_{4}\)T-COIs performs better than the investigated state-of-the-art approaches in terms of type and number of detected COIs.

在本文中，我们介绍了一种基于图文法属性的新方法来检测以社交网络形式表示的字段中的利益冲突（COI）。该方法包括专门研究Drewes等人的自适应星图语法（ASGG）。（Theor Comput Sci 411：3090–3109，2010）来表达我们称为\（K_4 \）型锦标赛图的子图类型，该子图对应于COI，无法由节点替换图语法生成。这种方法称为图语法和\（K_ {4} \）型基于锦标赛的方法，用于检测兴趣冲突\（（GGK_ {4} T-COIs）\），用于在通过社交网络代表的国际会议上接受的论文的审阅过程中检测COI。在此贡献中，所使用的图文法则的原理不是要考虑所有生成的语言，而是要考虑具有某些属性（对应于特殊图查询）的子图，这些子图可以标识表示COI的社交网络的各个部分。为了评估我们的命题的性能和效率，通过将其与文献中的并行方法进行比较，进行了实验。获得的结果表明，在检测到的COI的类型和数量方面，方法GG \（K_ {4} \） T-COI的性能优于所研究的最新方法。