Text this: Improved normalization and standardization techniques for higher purity in K-means clustering