The Application of Big Data Mining Prediction Based on Improved K-Means Algorithm
Yuchen Qiao, Yunlu Li, Xiaotian Lv · 2019
In order to solve the problem of low efficiency of K-Means algorithm in processing the data mining prediction problem of big data with more attributes, an annual income prediction method of residents based on improved K-Means algorithm is proposed. The improved K-Means algorithm combines the principal component analysis method with the traditional K-Means algorithm. After reducing the dimensionality of various data attributes, the data are classified with K-Means algorithm. The research makes use of 1994 U.S. census database and conducts a contrastive analysis of the two algorithms. The results show that the prediction accuracy has been significantly improved by 13.3313%, from 53.1016% to 66.4329%. It is clear the improved algorithm can effectively improve the accuracy of clustering and annual income prediction.