摘要
由于高校毕业生就业率与多种因素相关,具有复杂的变化规律,当前高校毕业生就业率预测模型存在一定的不足,如与实际值间的偏差大,建模时间长等,为了减少高校毕业生就业率预测误差,设计了一种基于混沌分析算法的高校毕业生就业率预测模型。首先,收集高校毕业生就业率的历史数据,结合历史数据的随机性、混沌性变化特点,采用混沌分析算法对历史数据的随机性、混沌性变化特点进行分析,重构高校毕业生就业率的历史数据;然后,引入当前最流行的数据挖掘技术——最小二乘支持向量机构建高校毕业生就业率的历史数据模型;最后,在相同平台上与当前经典高校毕业生就业率预测模型进行对比测试。结果表明,混沌分析算法的高校毕业生就业率预测值与实际值之间的偏差相当小,高校毕业生就业率预测精度超过94%,而经典模型的高校毕业生就业率预测精度处于90%左右,同时混沌分析算法减少了高校毕业生就业率预测的建模时间,可以满足现代高校毕业生就业率数据向大规模发展方向的要求。
For the college graduate employment rate is related with several factors and has complex changing rules,the current prediction model of the college graduate employment rate has some deficiencies,like large deviation from the actual values,long modeling duration,etc.In view of the above,a college graduate employment rate prediction model based on chaos analysis algorithm is designed to reduce the prediction error of the college graduate employment rate.The historical data of the college graduate employment rate are collected.The chaos analysis algorithm is used to analyze the characteristics of randomness and chaos of the historical data for the data reconstruction.And then,the most popular data mining technology named least square support vector machine is introduced for the model of the historical data of college graduate employment rate.Finally,comparative tests are performed on the same platform to compare the proposed model with the current classic employment rate prediction model.The results show that the deviation between the predicted value and the actual value of the college graduate employment rate by the chaos analysis algorithm is quite small,and the prediction accuracy of the college graduate employment rate exceeds 94%,while the prediction accuracy of the classic model is about 90%.At the same time,the chaos analysis algorithm can reduce the modeling duration of the prediction model,which meets the requirements of large⁃scale development of modern college graduate employment rate data.
作者
张英楠
ZHANG Yingnan(Jilin Agricultural University,Changchun 130000,China)
出处
《现代电子技术》
北大核心
2020年第21期101-105,共5页
Modern Electronics Technique
关键词
毕业生就业率
预测精度
数据挖掘技术
经典模型
随机性变化特点
混沌分析算法
graduate employment rate
prediction accuracy
data mining technology
classic model
random change characteristics
chaos analysis algorithm