In data mining, there are several algorithms that are often used in grouping data, including K-Means. However, this method still has several disadvantages, including the problem of the level of accuracy of the methods used to measure the similarities between the objects being compared. To overcome this problem, in this study a comparison was made between three methods (euclidean distance, manhattan distance, and minkowski distance) to determine the status of disparity in Teacher's needs in Tegal City. The results showed that of the three methods compared had a good level of accuracy, which is 84.47% (for euclidean distance), 83.85% (for manhattan distance), and 83.85% (for minkowski distance). In addition, this study also informs that there are still 6 (six) schools with conditions that are very poorly available for teachers (in the category of HIGH disparity labels) and need to get more attention, which is SMP Atmaja Wacana,

No abstract

In the big data era, the clustering of data or so-called clustering has attracted great interest or attention from researchers in conducting various studies, many grouping algorithms have been proposed in recent times. However, as technology evolves, data volumes continue to grow and data formats are increasingly varied, thus making massive data grouping into a huge and challenging task. To overcome this problem, various research related methods for data grouping have been done, among them is K-Means. However, this method still has some shortcomings, among them is the sensitivity issue in determining the value of cluster (K). In this paper we discuss the implementation of the rule-of-thumb approach and the normalization of data on the K-Means method to determine the number of clusters or K values dynamically in the data groupings. The results show that the implementation of the approach has a significant impact (related to time, number of iterations, and no outliers) in the data grouping.

Internet learning sebagai sebuah proses pembelajran berbasis elektronik yang digunakan sebagai media pembelajaran dierikan untuk para guru dan sekolah di SMA NU Ma’Arif Jatinegara Kota Tegal, melalui kegiatan pelatihan dilakukan sebagai upaya dalam upaya meningkatkan keterampilan dan pengetahuan guru dalam memahami dan membangun e-learning. Metode kegiatan yang digunakan adalah model pembelajaran partisipatif sehingga memberikan kebebasan kepada para peserta pelatihan untuk lebih memahami materi yang disampaikan. Media yang digunakan untuk membuat e-learning menggunakan Edmodo sebagai tools. Berdarsarkan hasil evaluasi terhadap kegiatan pelatihan tersebut memeprlihatkan hampir 80% peserta pelatiahan sudah mampu membuat media pembelajaran e-learning dengan baik.

Based on data from the Ministry of Health, the distribution ratio between health workers and patients in Indonesia is still not equal distributed. It influenced by the distribution of health human resources that are not in accordance with the ideal needs of health services. This results need to identify the profile of the availability of health human resources in Indonesia. In this study, an approach will be implemented to identify the profile of health human resources availability using K-Means Clustering with a combination of pillar algorithms in optimizing the selection of the initial cluster centroid. Chi-square analysis is used to determine the disparity in the needs of health human resources with the conditions of the availability of health human resources in the Central Java region. The data collection method used in this research is the observation method, while the scientific method used in this research is the K-Means Clustering method. The results showed that the application has been generated can dynamically determine the health human resource cluster based on the disparity category of health human resource availability in the Central Java region. In addition, the labeling of the Pillar K-Means cluster based on the Chi-square test has a high degree of accuracy, namely 80%.