Abstract

Reviewing is the most important step in the quality accreditation of scientific works, requires the professional expertise of the reviewer as well as there is no conflict of interest in the evaluation process. However, we also acknowledge that reviewers have limited knowledge, experience and opinions about the work of others, so they may misinterpret the author’s point of view, leading to the rejection of excellent scientific work or a potentially successful project proposal. Manually selecting reviewers can lead to bias and time-consuming. To solve these problems, we have developed a support system for selecting a group of reviewers to evaluate a particular problem, such as a proposal or a research paper. Our support system consists of three main modules: data collection, reviewer identification and group prediction of reviewers. The Data Collection module collects data from a variety of sources to create the scientist profile database. The reviewer identification module recognizes reviewers on a specific topic. The reviewer prediction module provides a group of exp reviewers to evaluate a submitted paper or a proposal. Experiments on the DBLP computer science bibliography dataset showed that our system achieves better results in terms of accuracy in comparison to other methods.

摘要

评审是科学作品质量认证中最重要的一步，需要评审者的专业知识，并且在评审过程中不存在利益冲突。但是，我们也承认，审稿人对他人工作的了解，经验和意见有限，因此他们可能会误解作者的观点，从而导致拒绝发表出色的科学著作或可能成功的项目建议。手动选择审阅者会导致偏见和耗时。为了解决这些问题，我们开发了一种支持系统，用于选择一组审阅者来评估特定问题，例如提案或研究论文。我们的支持系统包括三个主要模块：数据收集，审稿人识别和审稿人分组预测。数据收集模块从各种来源收集数据，以创建科学家档案数据库。审阅者识别模块识别特定主题的审阅者。审阅者预测模块提供一组exp审阅者来评估已提交的论文或提案。对DBLP计算机科学书目数据集的实验表明，与其他方法相比，我们的系统在准确性方面取得了更好的结果。