You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Avinash Kumar edited this page Oct 16, 2018
·
1 revision
BST
Normal BSTs can be unbalanced. So, their find/insert/deletion complexity can go upto O(n) in worst case. So, there are variants which balance the tree after every insertion.
AVL Tree - Use rotations after insertion/deletion.
Red-Black Tree - Colors + rotations. If the tree will be seeing many insertions/deletions compared to searching, then Red-Black tree is preferred as compared to AVL Tree because Red-Black tree avoids rotations in some cases.
Splay Tree - The rationale behind splay trees is that the nodes which are searched more should be at the top of the tree. When a node is accessed, it is promoted to the root of the tree. So, in practical situations Splay trees perform better than AVL and Red-Black trees. The difference here is that splay trees change their structure after a Search operation too.