QDU big data analytics autumn 2019 review focused on
five big questions
------------------------------- exam ----- ----------------------------
1. association rules
association rules: support (support), confidence (confidence) - key!
Strong association rules: minimum support, minimum confidence!
Apriori algorithm: find frequent item sets
[support, confidence]!
2. Category: decision trees, Bayesian
3. Classification index: ROC curve, PR diagram, recall, precision,
focusing on: concepts, but also requires drawing
two categories:
1) Classification index:
2) Return indicators:
4. clustering: only test based division! ! ! That is the midpoint of the improved
k-means: step algorithm, advantages and disadvantages of
k-means optimization: k-mode (Hamming distance), kprototype () ,! ! ! k- central point / PAM (data points when the central point)
5. comprehensive title: Classification comprehensive title! ! !
