情报科学 ›› 2021, Vol. 39 ›› Issue (7): 83-90.

• 业务研究 • 上一篇    下一篇

基于k-means与神经网络机器学习算法的用户信息聚类及预测研究

  

  • 出版日期:2021-07-16 发布日期:2021-07-16

  • Online:2021-07-16 Published:2021-07-16

摘要: 【目的/意义】基于机器学习算法对信息进行聚类及预测引起了广泛关注,本文将以航空公司客户信息为对
象构建出k-means,BP神经网络模型,对航空用户进行聚类及预测,实现用户的精准营销。【方法/过程】首先,对航
空公司的客户信息进行预处理,并根据信息聚类和信息预测理论,构建出k-means客户聚类模型与BP神经网络的
流失预测模型。【结果/结论】实证结果表明,在聚类模型上,k-means算法将客户聚为五类,实现了不同价值客户的
差异化识别;在客户预测模型上,BP神经网络的准确性更高。【创新/局限】本次研究将LRFMC模型引入到用户聚
类模型的实验中,使得模型泛化能力上存在了一定的局限,但也为该问题的未来研究提供了新的方式。

Abstract: 【Purpose/significance】By deeply analyzing the relationship among knowledge, routines and dynamic capabilities, this
study reveals the mechanism of knowledge and routines on dynamic capability, and provides a new perspective for the development of
dynamic capabilities.【Method/process】By reviewing relevant literature, this study explores the effects of knowledge creation and rou?
tine change on the formation of dynamic capabilities based on the paradigm of“knowledge-routine-capability”. Specifically, this pa?
per presents the process of knowledge creation by integrating internal and external knowledge, and reveals the changing process of or?
ganizational routines.【Result/conclusion】In the dynamic environment, knowledge creation and routine change are the crucial factors
influencing dynamic capabilities. Knowledge creation provides new and necessary knowledge for routine change. The update of dynam?
ic capabilities highly depends on the formation of new routines.【Innovation/limitation】This study analyzes the mechanism of knowl?
edge and routines on the formation of dynamic capabilities, and constructs a theoretical driving model, which still needs to be further
explored and tested in the future.