Hands-On Unsupervised Learning with Python

Chapter 1

1. Unsupervised learning can be applied independently from supervised approaches, because its goal is different. If a problem requires a supervised approach, often unsupervised learning cannot be employed as an alternative solution. In general, unsupervised methods try to extract pieces of information from a dataset (for example, clustering) without any external hint (such as the prediction error). Conversely, supervised methods require hints in order to correct their parameters.
2. As the goal is finding the causes of the trend, it's necessary to perform a diagnostic analysis.
3. No; the likelihood of n independent samples being drawn from the same distribution is obtained as a product of the single probabilities (see question 4 for the ...