Machine Learning notes

Supervised Machine Learning (providing labels)

Regression v.s. Classification:
one for the continuous data, the other for the discrete data.

LintCode 897. Island City - Weekly9

http://www.lintcode.com/en/problem/island-city/

You could use UnionFind, I just don't bother doing it.

As the islands are not dynamic UnionFind is unnecessary.

Just use plain old DFS, trigger it with a simple scan for cities that's not visited, then paint all visited cities with 2(or maybe not, it doesn't matter now).

Then count the number of triggers.

Boom!

LintCode 898. Leftmost One - Weekly9

http://www.lintcode.com/en/problem/leftmost-one/

An interesting questions, i find it sufficient applying binary search.

LintCode 895. Friend Request - Weekly9

http://www.lintcode.com/en/problem/friend-request/

It's a stupid question. Don't bother optimize with sorting or anything.

Just brute force is more than sufficient.

Nothing more to say...sigh...

LintCode 896. PrimeProduct - Weekly9

http://www.lintcode.com/en/problem/prime-product/

it's a tricky question if you are trying to if-else for 2 to 9 for-loops.

but we can simply use the DFS search.

Then eliminate duplication, and apply ordering.

LeetCode 794. Valid Tic-Tac-Toe State - Weekly Contest 74

https://leetcode.com/contest/weekly-contest-74/problems/valid-tic-tac-toe-state/




It's a simulation, just be patient and careful.

think of all the possibilities. Classified them. use | stand for the column strike, use - for the row strike.

use the \ and / for the diagonals.