An Improved Hybrid Algorithm for Web Usage Mining
Web mining is the application of data mining techniques to gather useful information from the World Wide Web. The rapid increase in digital use makes web usage mining (a subtype of web mining) important. To tackle the issues in web usage mining, we introduce a combination of hierarchical user emotion analysis and a self-organizing mapping algorithm in the training and testing of a recommended system. This method identifies the least dissimilar element, which will not last, and prefers the highest priority element in the cluster. The quality of the proposed system is evaluated in terms of entropy, purity, and Davies-Bouldin index. The proposed method is compared with various traditional clustering approaches such as ant colony clustering, k-means clustering, and genetic algorithm. The experimental results show that our proposed system provides 40% better quality when compared with traditional clustering approaches.
KeywordsData mining Web mining Web usage mining Recommended system User profiles
- 11.A. Abraham, V. Ramos, Web usage mining using artificial ant colony clustering and linear genetic programming. Congr. Evol. Comput. 2, 1384–1391 (2003)Google Scholar
- 13.P.N. Tan, M. Stenbach, V. Kumar, Introduction to Data Mining (Addison Wesley, Boston, 2005), pp. 1–6Google Scholar
- 14.D.L. Davies, D.W. Bouldin, Cluster separation measure. IEEE Trans. Pattern Anal. Mach. Intell. 1, 95–104 (1979)Google Scholar
- 16.The Jester Dataset. https://grouplens.org/datasets/jester/. 07 / 07/ 2001