Include this content: On the other hand, Z-score standardization transforms the data to have a mean of 0 and a standard deviation of 1. Using the same dataset (2, 4, 6, 8, 10), the mean is 6 and the standard deviation is about 2.828. After applying Z=XZ = \frac{X - \mu}{\sigma}Z=X, the values become approximately -1.414, -0.707, 0, 0.707, and 1.414. This method is especially useful for algorithms that assume normally distributed data, like logistic regression, because it preserves the distribution shape while making features comparable (King, n.d.). From my experience, I noticed that choosing the right method depends a lot on the algorithm. For kNN, min-max normalization makes sense because distances matter. But if I were using logistic regression or PCA, Z-score standardization would probably be a better choice because it adjusts for different scales and variances while keeping the data distribution meaningful (Skand, 2017)