Identifying evolutionary approach for search result clustering
Shashi Mehrotra, Shruti Kohli · International Conference on Computing for Sustainable Global Development · 2016
The study aims to identify an evolutionary method to improve search result of a website, which can be achieved by clustering the search result. Clustering search results organizes the results into meaningful folders, which makes browsing easier. There are many classical and evolutionary clustering algorithms are in use such as K-means, genetic clustering algorithm. Each clustering method has its advantages and disadvantages. The paper focuses on evolution technique and explores various evolution algorithms used for clustering. The Evolution approach such as genetic algorithm can overcome the drawbacks of the classical algorithms. Our earlier experiments about traditional clustering provide efficiency of K-Means in terms of accuracy and speed. Combining evolutionary approach with classical algorithm improves the efficiency. By combining K-Means with an evolutionary algorithm, K-Means improves compactness between clusters by reducing the difference between the clusters, and an evolutionary algorithm overcomes the drawback of K-means; the problem of convergence to local minima and to define the number of clusters in the beginning. The paper performs a comparative analysis of some most common evolutionary methods.