K-Means is a clustering algorithm in which the data required for clustering needs to be in the main memory. But as the size of the data increases, more data is required to be in the main memory which in turn reduces the computation time.
- The main idea of Mini-Batch Kmeans is to use small random batches of fixed-size data to be put in the main memory.
- A new random sample from the dataset is used to update the cluster in each iteration until convergence.
- The value of the learning rate decreases with each iteration.
- The position of the cluster’s centroid is updated based on new sample data points.