Criminal acts that endanger public health have seriously threatened people's health and life. How to prevent such criminal acts from occurring has become the focus of attention from all walks of life. There are few studies on the prevention of crimes endangering public health, and the results are not satisfactory. With the rapid development of artificial intelligence technology, machine learning algorithms are widely used in various fields. Based on the background of the times, this paper applies machine learning algorithms to the prevention research of crimes endangering public health, aiming to improve the efficiency of crime prevention. First of all, this paper establishes a predictive criminal behavior model based on support vector machine and random forest algorithm, and uses the model to analyze its performance. This article takes a certain city in our province as a specific investigation object, collects relevant case data of criminal acts endangering public health in the city from January to October 2018, and predicts criminal behaviors, and compares them with the actual crimes data collected later. At the same time, a random questionnaire survey was conducted on the citizens of this city to analyze the factors leading to crimes that endanger public health and their enthusiasm for participating in legislation. The experimental results show that Lagrangian interpolation can make the data set more complete, with a standard deviation of 1.19; the crime prediction model based on support vector machine and random forest algorithm can basically predict the incidence of crime, and the trend of its predicted data is basically consistent with the trend of actual data; 48.32% of the people believe that imperfect laws and regulations are the main reason for the frequent occurrence of crimes endangering public health, but only 18% are willing to actively participate in relevant legislation. The above results show that the prediction model established by artificial intelligence algorithm can effectively predict criminal behaviors that endanger public health and provide reliable data for prevention.

危害公众健康的犯罪行为已经严重威胁到人民群众的健康和生命。如何杜绝此类犯罪行为的发生，成为社会各界关注的焦点。关于预防危害公共健康犯罪的研究较少，结果也不尽如人意。随着人工智能技术的飞速发展，机器学习算法被广泛应用于各个领域。本文立足于时代背景，将机器学习算法应用于危害公共卫生犯罪的预防研究，旨在提高预防犯罪的效率。首先，本文建立了基于支持向量机和随机森林算法的犯罪行为预测模型，并利用该模型对其性能进行分析。本文以我省某市为具体调查对象，收集该市2018年1月至2018年10月期间危害公共卫生犯罪行为的相关案件数据，并对犯罪行为进行预测，并与之后收集的实际犯罪数据进行对比。同时，对本市市民进行了随机问卷调查，分析了危害公共卫生犯罪的发生因素及其参与立法的积极性。实验结果表明，拉格朗日插值可以使数据集更加完整，标准差为1.19；基于支持向量机和随机森林算法的犯罪预测模型基本可以预测犯罪发生率，其预测数据的趋势与实际数据的趋势基本一致；48.32%的人认为法律法规不完善是危害公共卫生犯罪频发的主要原因，但只有18%的人愿意积极参与相关立法。上述结果表明，人工智能算法建立的预测模型可以有效预测危害公共健康的犯罪行为，为预防提供可靠数据。