Which of the following is the GREATEST risk associated with normalizing a data set before...

Data normalization involves scaling data (e.g., ensuring all values are between 0 and 1). If you normalize the entire dataset before splitting it, the " Training Set " will be influenced by information from the " Testing Set " (such as the global maximum and minimum values). This is a form of " Data Leakage. " According to the AAIA™ manual, this " indirect knowledge " makes the model ' s test performance appear much better than it actually is, leading to a false sense of security. The correct procedure is to split the data first , then calculate normalization parameters using only the training data and apply those parameters to the test data.