基于海量数据分析的食品药品安全检测系统

发布时间：2018-05-21 05:41

本文选题：食品药品安全检测 + 海量数据　；参考：《西安工业大学》2017年硕士论文

【摘要】：如今食品药品安全事故频发,究其原因:一是各个地区检验所发展差距巨大,数据独立存储,格式标准不一,造成数据孤岛,数据资源难以共享;二是相关检验机构没有深入挖掘相关数据的潜在价值,从事后处理转变到事前预警上来。针对这些问题,本文研究了基于模糊分类的食品药品安全评估算法,对海量食品药品检验数据格式标准化,通过分析计算得出风险评估等级,使食品药品的安全问题得到了有效监管和安全预警。论文的主要工作如下:第一,研究了基于模糊分类的食品药品安全评估算法。对相关数据建立了数学模型,计算了检测项综合评价隶属度,通过模式识别判定了检品的质量级别。归纳统计出连续批次各质量等级检品数量所占比例,对各个质量等级检品建立了数据可视化模型,比较次等品变化趋势,结合知识规则,得到综合风险评估结果。第二,开发了基于海量数据分析的食品药品安全检测系统,将云计算、Hadoop计算框架、HDFS分布式文件系统、Map/Reduce的计算框架模型和HBase数据库等相关技术综合应用到食品药品安全检测系统当中,实现了对海量检验数据格式标准化,并通过存储服务器的虚拟化以及数据资源的分布式存储,解决了底层存储设施的空间浪费、负载均衡、动态备份、动态迁移等问题,使得系统性能稳定,且具有较高的可靠性、可扩展性以及良好的交互界面。第三,实验以某发酵乳数据集进行了实验分析,数据集中有20个不同品牌发酵乳,每个品牌20个批次检品,每批次20个样品共计16.8万数据。通过分析计算得出了这20个品牌检品的风险等级。实验结果表明该算法对于食品药品质量风险评估是有效的。本文在所设计构建的基于海量数据分析食品药品安全检测系统,将海量、多样的检测数据进行高效存储,实现大规模复杂检测数据查询的快速和准确响应;通过对现有数据集的分析挖掘,评估检品风险等级,为食品药品安全预警提供数据支持,系统具有一定的实用价值。
[Abstract]:Nowadays, food and drug safety accidents occur frequently, the reasons are as follows: first, there is a huge gap in the development of inspection institutes in various regions, the data is stored independently and the format is different, resulting in isolated data islands and difficult to share data resources; The other is that the relevant inspection institutions have not excavated the potential value of the relevant data deeply and changed from the post-treatment to the pre-warning. Aiming at these problems, this paper studies the food and drug safety assessment algorithm based on fuzzy classification, standardizes the format of mass food and drug inspection data, and obtains the risk assessment level through analysis and calculation. Food and drug safety issues have been effectively monitored and early warning. The main work of this paper is as follows: first, the food and drug safety assessment algorithm based on fuzzy classification is studied. The mathematical model of the related data is established, the membership degree of the comprehensive evaluation of the test item is calculated, and the quality grade of the inspection product is determined by pattern recognition. The proportion of the quantity of each quality grade inspection product in successive batches is summed up and the data visualization model is established for each quality grade inspection product. The change trend of the inferior product is compared and the comprehensive risk assessment result is obtained by combining the knowledge rules. Secondly, a food and drug safety detection system based on mass data analysis is developed. In this paper, the computing framework of Hadoop computing framework, HDFS distributed file system, HBase database and other related technologies are comprehensively applied to the food and drug safety inspection system, and the format of mass inspection data is standardized. Through the virtualization of storage server and the distributed storage of data resources, the problems of space waste, load balance, dynamic backup, dynamic migration of the underlying storage facilities are solved, which makes the system performance stable and has high reliability. Extensibility and good interface. Third, the experiment was carried out with a fermented milk data set. There were 20 different brands of fermented milk, 20 batches of each brand, 20 samples per batch, a total of 168000 data. Through the analysis and calculation, the risk grade of the 20 brand inspection products is obtained. Experimental results show that the algorithm is effective for food and drug quality risk assessment. In this paper, the food and drug safety detection system based on mass data is designed and constructed, which stores the massive and diverse detection data efficiently, and realizes the fast and accurate response of large-scale complex detection data query. By analyzing and mining the existing data sets and evaluating the risk grade of the inspection products, the system has certain practical value to provide data support for food and drug safety early warning.
【学位授予单位】：西安工业大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TS207.3;TP311.13;TQ460.7

【相似文献】