摘要
机器人与机器学习每日新闻的新闻记者兼工作人员新闻编辑-关于机器学习的最新研究结果已经发表。根据NewsRx编辑对中国青岛的新闻报道，研究表明：“复杂的风险因素使地铁施工安全事故容易发生，事故类型多样。事故报告以文本形式记录了不同类型事故的详细信息。”我们的新闻记者从青岛理工大学的研究中获得了一句话：“然而，有效利用这种非结构化数据带来了巨大的挑战。文本挖掘™为应对这一挑战提供了可行的基础。”但相关研究在风险特征提取方面存在局限性,缺乏深入分析能力。针对现有研究的不足,为地铁建设领域关键风险因素的识别提供可行的策略,本文提出了一种基于TM和机器学习的贝叶斯网络相结合的综合模型。利用TM中的术语frequenc y-逆文档频率(TF-IDF)算法分别提取事故报告中的直接原因和间接原因，并利用TextRank算法补充各因素，然后根据是否考虑因素间条件独立性的假设，基于提取的因子和相应的攻击类型,分别构建了改进的朴素贝叶斯网络(NBN)和树增广朴素贝叶斯网络(TAN),进一步深入分析,最后将训练集划分为训练两个网络模型,并利用敏感性分析识别关键危险因素,以162份中国事故报告为例,进行了应用分析。结果表明，与改进的NBN(71.75%)相比，TAN在测试集中的平均准确率(79.62%)更高，并成功地从多个角度对不同事故类型危险因素的重要性进行了排序，同时获得了对我国地铁事故的一些新认识，为事故预防和控制决策提供了依据。
Abstract
By a News Reporter-Staff News Editor at Robotics & Machine Learning Daily News – Current study results on Machine Learning have be en published. According to news reporting out of Qingdao, People’s Republic of C hina, by NewsRx editors, research stated, “Complex risk factors make metro const ruction safety accidents prone to occur, and there are various types of accident s. Accident reports record detailed information about different types of acciden ts in text form.” Our news journalists obtained a quote from the research from the Qingdao Univers ity of Technology, “However, effectively utilizing such unstructured data presen ts a significant challenge. Text mining ™provides a viable foundation for addre ssing this challenge, but related studies have limitations in risk feature extra ction and lack of in-depth analysis capability. To address the deficiencies of e xisting studies and provide a feasible strategy for identifying key risk factors in the metro construction domain, this paper proposes an integrated model combi ning TM and machine learning-based Bayesian networks. Firstly, the term frequenc y-inverse document frequency (TF-IDF) algorithm in TM was used to separately ext ract the direct and indirect cause factors from the accident reports, with the m issing factors supplemented using the TextRank algorithm. Then, depending on the assumption of whether to consider the conditional independence between factors, an improved naive Bayesian network (NBN) and a tree-augmented naive Bayesian ne twork (TAN) were built based on the extracted factors and the corresponding acci dent types, respectively, for further in-depth analysis. Finally, the training s et was divided to train the two network models, and sensitivity analysis was use d to identify the key risk factors. Using 162 accident reports from China as an application example, the results showed that TAN exhibited a higher average accu racy (79.62 %) in the test set compared with the improved NBN (71.75 %), and the importance of risk factors for different accident types was successfully ranked from multiple perspectives using TAN. Meanwhile, some n ew insights into metro accidents in China were obtained, which can support decis ion-making for accident prevention and control.”
NETL
NSTL