K-Means-Clustering and Self-Organizing Maps Comparison to Classify the Electricity Generation in Kurdistan

Most fundamental unsupervised learning challenge is clustering. It is the process of grouping elements into groups whose components are “similar but not identical” to objects in other groups. It is a technique for classifying subjects into clusters of similar components; we attempt to identify related items based on their properties. We develop homogenous groupings, or clusters, that are distinct from one another. A Self-Organizing Maps (SOM) network's output, from the other hand, does not really provide groups of locations on the map. The SOM map is now arranged so that the network size on the map corresponds to the number of clusters required. Clustering aim is to find the intrinsic grouping in a group of data set. In this study, two methods for generating power in the Kurdistan region were compared by using methods (Self Organize Maps and K-Mean Clustering) to classify the months the data was in (Iraq-Kurdistan Regional Government Ministry of Electricity General Directorate of Control &Communication Kurdistan Dispatch Control Centre) from 2006 to 2019. The research found that the months were divided into three groups, and that the Self-Organizing Maps algorithm outperformed the K-means strategy in getting the results with the lowest distance in a cluster. As a result, the SOM is the best and most accurate method for data classification, which supports its implementation. The classification process can also be adjusted by the neural network to account for changes in the power generated by electric patterns over time.