This paper aims at exploring the ways of artificial intelligence to identify and utilize the big data of journals, and adjusting the ways to review journal manuscripts by the collaborative management model of artificial intelligence editing and traditional editing, in an effort to improve the efficiency of initial evaluation. By studying the behavior of authors submitting a paper to Journal of Sichuan Forestry Science and Technology (JSFST), the author tries to find out any unforced regularity in the submission behavior and the relationship between the quality of articles. The data of papers submitted to the submission system of JSFST from November 28, 2017 to November 28, 2019 (i.e. over 2 years, totaling 731 days) were analyzed, with chi-square test available to maintain the validity of the findings. In the context of entropy, the thermodynamic relation is proposed, and the entropy distance is used to measure the disorder of the submission process. The numbers and percentages of accepted and rejected papers as well as the ratio of accepted to rejected papers are compared, and the submission day is checked. The author proposes a new review strategy with particular emphasis on weekly submission day, thus improving effectively the efficiency of initial evaluation. As a new data utilization pattern of deep learning with big data by artificial intelligence, this strategy provides a new way for editors to rapidly screen articles in the era of big data and artificial intelligence, which will inject more vitality into the publishing process and make it easier to find high-quality articles.

本文旨在探索人工智能识别和利用期刊大数据的方式，并通过人工智能编辑与传统编辑的协同管理模型来调整期刊稿件的审阅方式，以期提高初稿的效率。评价。通过研究作者的行为向“四川林业科技”投稿（JSFST），作者试图找出提交行为中的任何非强制性规律以及文章质量之间的关系。分析了2017年11月28日至2019年11月28日提交给JSFST提交系统的论文数据（即2年以上，共731天），并进行卡方检验以保持研究结果的有效性。在熵的背景下，提出了热力学关系，并利用熵距离来衡量提交过程的无序性。比较接受和拒绝的论文的数量和百分比，以及接受与拒绝的论文的比例，并检查提交日期。作者提出了一种新的审阅策略，特别侧重于每周提交日，从而有效地提高了初始评估的效率。