Implementation of K-Medoid and Hierachical Clustering for Strengthening Job Satisfaction on The Teacher Profession

Education is a very important element in social life to support the formation of individual character. The implementation of data mining in the education sector shows an increasing trend. The purpose of this study is to implement strengthening job satisfaction for the teaching profession using the K-medoid and Hierarchical Clustering method. Cluster analysis was carried out on five main factors including salary, promotion, supervisor's supervision, relationships with colleagues and the teaching profession. Tests were carried out on 136 respondents and 39 questions devided into 5 factors. The trial of determining the number of clusters was carried out on 3 trials ($\mathrm{K}=2, \mathrm{K}=3$ and $\mathrm{K}=4$). The optimal cluster value is obtained at $\mathrm{K}=3$ with a Davies Bouldin index (DBI) of 2,823 for the K-Medoid method and $\mathrm{K}=4$ with a DBI value of 1,415 for the Hierarchical method. The cluster results show that the priority of strengthening teacher professional satisfaction needs to be increased on the salary indicator by 50% for the K-Medoid method and 65% for the Hierarchical method. While the results of the percentage of priorities that need to be maintained are the indicators of relationships with colleagues by 10% in the K-Medoid method and 5% in the Hierarchical method.