Clustering Data Remunerasi PNS Menggunakan Metode K-Means Clustering Dan Local Outlier Factor

Remuneration is a reward for services performed by employees. Remuneration is given to employees who have good performance. Remuneration can be given by an agency if it has implemented the Financial Management of Public Service Bodies (PK-BLU). The problem with remuneration is that the validation carried out by the direct supervisor of the employee concerned is still doubtful. Based on this, we need a system that can detect outlier data from the remuneration data of Civil Servants and classify the data using data mining techniques. The algorithm for finding outlier data is the Local Outlier Factor (LOF) algorithm and the algorithm that can be used to do clustering is a k-means clustering algorithm. The K-means algorithm has problems in determining the optimal number of clusters. Problems with the K-means method can be solved using the elbow method. Determination of this method is seen from the Sum Square Error (SSE) graph of several cluster numbers. This study aims to classify the remuneration data of civil servants by using the k-means clustering method with improvisation at the pre-processing stage and determining the optimal number of clusters. Local Outlier Factor method with a MinPts value of 150 can detect the most outlier data with 162 data outliers detected or 22.98%. The optimal number of clusters with the elbow method is 4 clusters with a Silhoutte value of 0.542, Dunn of 0.040 and Purity of 0.89