基于箱形图和隔离森林的施工人次数据处理与预测研究

针对工程项目进度和施工人员进场记录数据存在异常值的问题，提出一种基于箱形图和隔离森林的施工人次数据处理方法。将历史工程项目进行分类并计算各工程项目每人次的效率值，采用箱形图法剔除效率值异常的样本；结合各类型工程项目投资和投入施工人次的数据可视化情况，通过隔离森林隔离出属性值异常的数据点；采用剔除异常值后的样本作为线性回归预测模型的学习样本，使用测试集进行预测精度测试。并以某施工单位的实测数据为具体对象，根据所提算例结果显示，该方法能实现异常值的定位与分离，有效提高施工人次预测的精确度。

Data Processing and Forecast Analysis of Construction Workers Based onBoxplot Method and Isolation Forest

In order to solve the problem of outliers in the data of project progress and construction staff attendance，this paperproposes an abnormal data processing method based on boxplot and isolation forest. Firstly，efficiency value of each project iscalculated and the abnormal efficiency value is removed by the boxplot method. Then，combining with the visualization of theinvestment and construction workers value，outliers are isolated by isolation forests. Finally，the sample after removing the outliers isused as training sample for each type of linear regression prediction model，and the test samples are used for accuracy testing. Basedon the specific measurement data，simulation results show that the method can determine the location and separate the abnormal datacorrectly，which establishes the foundations for enhancing the precise of forecasting.