基于K-Means聚类的多元回归拟合定价模型

Multiple Regression Fitting Pricing Model Based on K-Means Clustering

摘要: 近年来，基于移动互联网的自助式劳务众包平台日渐火爆，用户可通过拍照做任务赚取酬金，其相比传统的市场调查方式可以大大节省调查成本，平台中的任务定价是其核心要素。若因为定价不合理，则用户人数减少，影响平台任务的正常运行，导致平台竞争力减弱。针对如何合理定价这一问题，本文采用“K-means聚类”作为主要研究手段，运用多元回归方程法，考虑多种因素的影响并归一化分析，取离散度、集中度最优的因素作为主要影响因子，建立多元回归方程，以利润为指标，找出各因素与价格最优的关系，评判本多元回归拟合定价模型的优劣。

Abstract: In recent years, the self-service labor crowd-sourcing platform based on mobile internet has be-come increasingly popular. Users can earn remuneration by taking photos to do tasks. Compared with traditional market research methods, it can greatly save the cost of investigation. The task pricing in the platform is its core element. If the pricing is not reasonable, the number of users will decrease, which will affect the normal operation of the platform tasks and weaken the competitive-ness of the platform. To solve the problem of how to reasonably price, this paper uses “K-means clustering” as the main research means, uses multiple regression equation method, considers the influence of multiple factors and normalizes the analysis, takes the factors with the best dispersion and concentration as the main influencing factors, establishes multiple regression equation, takes profit as the index, finds out the relationship between each factor and the best price, and judges the advantages and disadvantages of this multiple regression fitting pricing model.