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

Clustering and Prediction of User Information Based on k-means and Neural Network Machine Learning Algorithm

【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。

【Purpose/significance】Based on the machine learning algorithm, the k-means, bp neural network model is constructed based on the airline customer information, and the aviation user information is clustered and users.【Method/process】Firstly, the customer information of airlines is preprocessed, and the k-means customer clustering subdivision model and the loss prediction model of bp neural network are constructed according to the information clustering and information prediction theory.【Result/conclusion】The empirical results show that the k-means algorithm gathers the customer information into five categories in the clustering model, and realizes the differential recognition of customers with different value, and the accuracy of bp neural network is higher in the customer prediction model.【Innovation/limitation】This study introduces the LRFMC model into the experiment of user clustering model, which makes the generalization ability of the model limited, but also provides a new way for the future research of this problem.