摘要
拍照赚钱是用户会员通过完成需要拍照的任务,赚取对任务所标定的酬金。基于合理的任务定价提高任务的完成率的目的。采用以会员能力与信誉度的综合量化值作为基,利用高斯分布函数特有的性质来计算出权重,插值计算出等价高斯效能分布图的方法,提出任务位置在等价高斯效能分布图中的高斯效能值与定价的乘积为完成预期值,结合经验数据中完成任务的完成预期值给出了定价方案。为进一步优化模型提高任务完成率,本文提出了对会员进行离群点过滤与K均值聚类将任务打包方案。结合广东深圳的数据进行实验,结果表明本文的定价方案将原有任务完成率提高了28.62%。
Making money by taking pictures is to earn a fee for the members by completing the tasks that need to take picture.Based on the purpose of reasonable task pricing to improve the completion rate of tasks.We adopt the following methods:a comprehensive quantitative value of membership competence and credibility is proposed as the basis,calculate weights by using the special properties of gauss distribution function,calculate equivalent gauss effictiveness distribution by interpolation,propose that the product of the gauss efficiency value location in equivalent gauss efficiency graph and the price of the task is the expected completion value.Base on the expected value of accomplishing tasks in empirical data,a pricing scheme is given.For optimizing model and improving task completion rate,in this paper,propose an outlier filtering for members and a k means clustering scheme to package tasks.Experiments with the data from Shenzhen,Guangdong,the results show that under the premise of not exceeding the total budget,the pricing scheme in this paper improves the original task completion rate by 28.62%.
作者
张永亮
齐志博
申懿
ZHANG Yong-liang;QI Zhi-bo;SHEN Yi(15th Research Institute of China Electronics Technology Group Corporation,Beijing 100083,China;Shenyang Aircraft Design and Research Institute of Aviation Industry,Shenyang 110000,China)
出处
《电子设计工程》
2020年第5期40-46,共7页
Electronic Design Engineering
关键词
高斯分布函数
高斯效能分布图
完成预期值
离群点过滤
K均值聚类
Gauss distribution function
Gauss efficiency map
complete expected value
outlier filtering
K means clustering
