The Application of Elastic Network Algorithm Based on Weighted Property in Clustering Analysis

Abstract

The rapid development of internet technology has brought convenience to people’s lives, increasingly making their work and personal activities inseparable from the internet. However, the various types of information data in the Internet are very large, and accurately obtaining useful information from them has become particularly important. Therefore, clustering analysis technology is widely applied in various fields. The elastic network algorithm is vulnerable to noise and other factors in clustering analysis, resulting in low accuracy of clustering results. To solve this problem, this study proposes an elastic network algorithm based on weighted properties for clustering analysis. After comparative experiments, the results showed that the accuracy of the proposed algorithm was 98%. The area under the precision recall curve was 0.86. The maximum F-value was 0.93. The area under the working characteristic curve of the subject was 0.94. The precision ultimately stabilized at 97%. All values were higher than the comparison algorithm. In the experiment, the running time of the algorithm was 11 seconds and the average error was 7%, both of which were lower than the comparison algorithm. When the algorithm was affected by Statistical dispersion and threshold parameters, the accuracy of the proposed algorithm was 71%, 69% and 67% respectively, which were better than the comparison algorithm. In conclusion, the proposed elastic network algorithm based on the weighted characteristics has high accuracy, strong stability and advantages in clustering.

Keywords

• Elastic network
• cluster analysis
• weighted characteristics
• maximum entropy