Development of a novel hybrid algorithm for data clustering

In this thesis, Whale Optimization Algorithm (WOA), a global optimization algorithm developed by inspiration from hunting behaviors of humpback whales, has been improved with chaos maps and then hybridized with Fuzzy C-Means (FCM) algorithm which is frequently used in data clustering problems and chaotic BOA algorithms are proposed. The performances of the proposed algorithms are evaluated with mean benchmark function, standard deviation and Wilcoxon Sign Rank Test at 0,05 significance level and tested with 13 different benchmark functions. Each of the proposed chaotic WOA algorithms is compared with the WOA algorithm, and the performance improvement is shown both statistically and graphically. Then, the proposed chaotic WOA algorithms are integrated with the FCM algorithm and hybrid data clustering algorithms are proposed. The clustering performances of hybrid data clustering algorithms measured with objective function, Rand Index and Adjust Rand Index values are compared with the FCM algorithm for 7 different data sets selected from the UCI Repository database. In addition, new hybrid clustering algorithms are improved by using Chebyshev distance function instead of Euclidean distance in FCM algorithm to increase data clustering performance of proposed hybrid algorithms. All proposed algorithms are compared with each other and FCM algorithm, and it is observed that they can cluster data better than FCM algorithm. As a result, it has been seen that chaos functions improve the optimization performance of WOA algorithm, integrating chaos-based WOA algorithms with FCM algorithm improves disadvantages of FCM algorithm, changing distance function increases clustering performance of algorithms.