摘要
文章研究了K-means聚类分析下的"拍照赚钱"任务定价方案设计。"拍照赚钱"是移动互联网下的一种自助式服务模式。用户注册为会员,从APP上领取需要拍照的任务,赚取APP对任务所标定的酬金。对任务数据进行预处理,去除不合理数据,之后将任务地点经纬度转换为实际距离,建立一个包含所有任务的区域。对区域中的任务,计算以任务为中心,半径10 km范围内所有会员的归一化后的配额、信誉度、距任务点的平均距离。对3组数据及其任务对应的定价进行多元线性回归分析。通过K-means聚类分析,将任务打包,得到80个任务包的质心。将得到的80个任务包视为80个新任务。其质心视为新任务地点。距离质心最近的任务价格和任务包的任务数的乘积视为任务包价格。获得任务包周围会员配额、会员信誉度及会员距任务地点的平均距离后,将价格与该3个因素进行多元线性回归,可得任务打包情况下的价格函数。
This paper studies the design of pricing scheme for 'taking photos to make money' task under K-means cluster analysis.The 'taking photos to make money' is a self-service mode under the mobile Internet. The user is registered as a member, and receives the task that needs to be photographed from the APP, and earns the rewards that the APP has calibrated for the task. After preprocessing the task data to remove the unreasonable data, the longitude and latitude of the task site were converted into the actual distance, and a region containing all the tasks was established. For the tasks in the area, calculate the normalized quota, reputation and average distance from the task point of all members within a 10-kilometer radius with the task as the center. The pricing of three groups of data and their corresponding tasks were analyzed by multiple linear regression. Through K-means clustering analysis, the task is packaged to obtain the centroid of 80 task packages. Treat the 80 task packages you get as 80 new tasks. Its center of mass is regarded as the new mission site.The product of the task price nearest to the center of mass and the number of tasks in the task package is regarded as the task package price. After obtaining the quota of members around the task package, the member’s reputation and the average distance from the task site,multiple linear regression was conducted between the price and the three factors to obtain the price function under the task package.
作者
李昊哲
Li Haozhe(Jiangnan University,Wuxi 214122,China)
出处
《无线互联科技》
2019年第4期64-65,共2页
Wireless Internet Technology
关键词
拍照赚钱
K-means聚类分析
多元线性回归
定价
任务包
taking photos to make money
K-means cluster analysis
multiple linear regression
pricing
task packages