室内公共场所空气质量指标的对数线性模型分析

Study on the Relationship of Classified Air Quality Indexes in Public Places by Hierarchical Loglinear Model

[目的 ]对公共场所多种空气质量分类指标间的关系 ,用多维列联表来表达 ;分析气温 (TEM)、负离子浓度(NI)、二氧化碳 (CO2 )含量对空气中的细菌总数 (BAC)的影响 ;用对数线性模型来评价效应的主次和估计交互作用的大小。[方法 ]收集 5家歌舞厅 6 6个监测点的上述 4个空气质量指标的数据 ,以 SPSS10 .0 for Windows建立数据库并进行数据处理。用对数线性模型分析。 [结果 ]通过其最佳模型 ,发现影响空气中细菌总数多少的主要因素是二氧化碳与负离子浓度 ,其中 CO2 * BAC交互作用最大 ,参数估计为 +0 .46 73;其次是 CO2 * NI,参数估计为 - 0 .42 46。 [结论 ]二氧化碳含量与负离子浓度均会影响空气中的细菌总数 ,前者为正向影响 ,后者为反向影响。增加通风以减少二氧化碳蓄积并增设负离子发生器 ,可减少室内公共场所空气中的细菌总数 ,改善空气质量。对数线性模型检验法可更有效地分析这种多个分类指标的计数资料间的关系 (高维列联表 )

Objective] To study on the relationship of multiple classified air quality indexes of public places and analyze the influence of 3 indexes (temperature, carbon dioxide and negative ion) on total number of bacteria. The influence order and value of interactive effects are analyzed. [Methods] The classified multiple air indexes of 66 points in Fuzhou are shown in a multi dimensional contingency table. Fitting a hierarchical loglinear model (saturated) instead of traditional χ 2 test for the table. The data were collected and analyzed by HILOGLINEAR command in SPSS 10 0.5 for Windows. [Results] The most important factors of influencing bacteria number are carbon dioxide and negative ion. Among them, the largest iterative effect index in the final model best fitted is “carbon dioxide * bacteria”. The estimated parameter is +0.4673; The second one is “carbon dioxide * negative ion”,-0 4246. [Conclusions] Both carbon dioxide and negative ion can influence the number of bacteria. Increasing the ventilation can decrease the accumulation of carbon dioxide; Raising the concentration of negative ion can also improve the air quality. Hierarchical loglinear model can analyze multi dimensional contingency table of multi index, classified variables more systematically than χ 2 test.