The datasets for short-term modeling and long-term modeling are clustered respectively throughstandard clustering techniques like K-means [33]. The K-means algorithm randomly selects K centers. Each data point then chooses to join the cluster of a specific center if the distance between the data point and the center is the shortest (comparing to the distances to other cluster centers). It iteratively updates the cluster centers and members of each cluster until it is converged. Manhattan distance is used here to measure the similarity between data points.
đang được dịch, vui lòng đợi..
