Arithmetic Slices II - Subsequence LT446
A sequence of numbers is called arithmetic if it consists of at least three elements and if the difference between any two consecutive elements is the same.
For example, these are arithmetic sequences:
1, 3, 5, 7, 9
7, 7, 7, 7
3, -1, -5, -9
The following sequence is not arithmetic.
1, 1, 2, 5, 7
A zero-indexed array A consisting of N numbers is given. A subsequence slice of that array is any sequence of integers (P0, P1, ..., Pk) such that 0 ≤ P0 < P1 < ... < Pk < N.
A subsequence slice (P0, P1, ..., Pk) of array A is called arithmetic if the sequence A[P0], A[P1], ..., A[Pk-1], A[Pk] is arithmetic. In particular, this means that k ≥ 2.
The function should return the number of arithmetic subsequence slices in the array A.
The input contains N integers. Every integer is in the range of -231 and 231-1 and 0 ≤ N ≤ 1000. The output is guaranteed to be less than 231-1.
Example:
Input: [2, 4, 6, 8, 10] Output: 7 Explanation:
All arithmetic subsequence slices are:
[2,4,6]
[4,6,8]
[6,8,10]
[2,4,6,8]
[4,6,8,10]
[2,4,6,8,10]
[2,6,10]
Idea 1. BruteForce. 穷举所有符合条件的序列,序列其实是数组的subsets, 用DepthFirst Search穷举subsets.
Time complexity: O(2^n) For each element in the array, it can be put in or outside of the subsequence, two choices for each element.
Space complexity: stack depth O(n)
class Solution {
private boolean isArithmeticSequence(int[] A, int currDep) {
if(currDep < 3) {
return false;
}
long diff = (long)A[1] - A[0];
for(int i = 2; i < currDep; ++i) {
if(diff != (long)A[i] - A[i-1]) {
return false;
}
}
return true;
}
private void helper(int[] A, int depth, int[] path, int pathPos, int[] count) {
if(depth == A.length) {
if(isArithmeticSequence(path, pathPos)) {
++count[0];
}
return;
}
helper(A, depth+1, path, pathPos, count);
path[pathPos] = A[depth];
helper(A, depth+1, path, pathPos+1, count);
}
public int numberOfArithmeticSlices(int[] A) {
int[] path = new int[A.length];
int[] count = new int[1];
helper(A, 0, path, 0, count);
return count[0];
}
}
Note: 1. reset the change on the current depth before backtracking to the previous depth. List implementation is more obvious, as array just keep the int index pathPos unchanged.
2. Overflow, change int to long to filter out invalid cases, as there is no valid arithmetic subsequence slice that can have difference out of the Integer value range.
class Solution {
private boolean isArithmeticSequence(List<Integer> curr) {
if(curr.size() < 3) {
return false;
}
long diff = (long)curr.get(1) - curr.get(0);
for(int i = 2; i < curr.size(); ++i) {
if(diff != (long)curr.get(i) - curr.get(i-1)) {
return false;
}
}
return true;
}
private void helper(int[] A, int depth, List<Integer> curr, int[] count) {
if(depth == A.length) {
if(isArithmeticSequence(curr)) {
++count[0];
}
return;
}
helper(A, depth+1, curr, count); // not put A[depth] in the subsequence
curr.add(A[depth]);
helper(A, depth+1, curr, count); // put A[depth] in the subsequence
curr.remove(curr.size()-1); // reset before backtracking
}
public int numberOfArithmeticSlices(int[] A) {
int[] count = new int[1];
helper(A, 0, new ArrayList<>(), count);
return count[0];
}
}
python:
class Solution:
def isArithmetic(self, curr: List[int]) -> bool:
if(len(curr) < 3):
return False; diff = float(curr[1]) - curr[0]
for i in range(2, len(curr)):
if diff != float(curr[i]) - curr[i-1]:
return False; return True def helper(self, A: List[int], depth: int, curr: List[int], count: List[int]) -> None :
if depth == len(A):
if self.isArithmetic(curr):
count[0] += 1 return self.helper(A, depth+1, curr, count)
curr.append(A[depth])
self.helper(A, depth+1, curr, count)
curr.pop() def numberOfArithmeticSlices(self, A: List[int]) -> int:
count = [0]
self.helper(A, 0, [], count)
return count[0]
Idea 2: Dynamic programming, similar to Arithmetic Slices LT413, how to extend from solution to nums[0...i] to nums[0..i, i+1]? LT413的sequence要求是连续的,只需要检查能否延续前一位为结尾的序列,一维的关系:dp(i) = dp(i-1) + 1; 而这一题可以跳过前面的数,延续前面任何以nums[j]结尾的满足条件的序列(0 <j <i, diff(nums[k, j]) = nums[i] - nums[j]),需要加入序列的差d来表达关系,用dp(i, d)表示以nums[i]结尾,序列差为d的序列个数,
dp(i, d) = sum(dp(j, d) + 1)
序列要求是三位数的长度,如果以3位数为base case这个并不好计算,如果放松一下条件2位数算作wealy arithmetic sequence, 上面的公式依然成立,2位数的base case也好计算,
dp(i, nums[i]-nums[j]) = 1 for any pair j, i, 0 <= j < i
我们来走一下例子:[1, 1, 2, 3, 4, 5]
i = 0, dp(0, d) = 0
i = 1, j = 0, diff = 1 - 1 = 0, dp(1, 0) = 1, sequence: [1, 1]
i = 2, j = 0, diff = 2 - 1 = 1, dp(2, 1) = 1; j = 1, diff = 2 - 1 = 1, dp(2, 1) = 1 + 1 = 2 sequence: [1, 2], [1, 2]
i = 3, j = 0, diff = 2, dp(3, 2) = 1; j = 1, diff = 2, dp(3, 2) = 2; j = 2, diff = 1, dp(3, 1) = dp(2, 1) + 1 = 3, sequence: [1, 3], [1, 3], [1, 2, 3], [1, 2, 3], [2, 3]
i = 4, j = 0, diff = 3, dp(4, 3) = 1; j = 1, diff = 3, dp(4, 3) = 2; j = 2, diff = 2, dp(4, 2) = 1; j = 3, dp(4, 1) = dp(3, 1) + 1 = 4, sequence: [1, 4], [1, 4], [2, 4], [1, 2, 3, 4], [1, 2, 3, 4], [2, 3, 4], [3, 4]
i = 5, j = 0, diff = 4, dp(5, 4) = 1, sequence[1, 5]; j = 1, diff = 4, dp(5, 4) = 2, sequence [1, 5] [1, 5]; j = 2, diff = 3, dp(5, 3) = 1; j = 3, diff = 2, dp(3, 2) = 2, dp(5, 2) = dp(3, 2) + 1 = 3, sequence [1, 3, 5], [1, 3, 5], [3, 5]; j = 4, dp(5, 1) = dp(4, 1) = 5, sequence [1, 2, 3, 4, 5], [1, 2, 3, 4, 5], [2, 3, 4, 5], [3, 4, 5], [4, 5]
从例子可以看出来符合至少3位数的序列个数其实取决于前面sequence个数dp(j, d), 公式中的+1是pair (nums[j], nums[i])2位数的序列,总结公式如下:
dp(i, d) = sum(dp(j, d) + 1)
dp(i, nums[i]-nums[j]) = 1 for any pair j, i, 0 <= j < i
result(3位数的序列个数) = sum(dp(j, d))
由于 d是unbounded可正可负,一般dynamic programming使用二维数组做memory就不能用了,而用array of map, dp(i).get(d) = dp(i, d)
Time complexity: O(n2)
Space complexity: O(n2)
class Solution {
public int numberOfArithmeticSlices(int[] A) {
int result = 0;
List<Map<Integer, Integer> > dp = new ArrayList();
for(int i = 0; i < A.length; ++i) {
dp.add(new HashMap());
for(int j = 0; j < i; ++j) {
long delta = (long)A[i] - A[j];
if(delta < Integer.MIN_VALUE || delta > Integer.MAX_VALUE) {
continue;
}
int diff = (int) delta;
int prev = dp.get(j).getOrDefault(diff, 0);
int curr = dp.get(i).getOrDefault(diff, 0);
dp.get(i).put(diff, curr + prev + 1);
result += prev;
}
}
return result;
}
}
array of map
class Solution {
public int numberOfArithmeticSlices(int[] A) {
int result = 0;
Map<Integer, Integer>[] dp = new Map[A.length];
for(int i = 0; i < A.length; ++i) {
dp[i] = new HashMap<>();
for(int j = 0; j < i; ++j) {
long delta = (long)A[i] - A[j];
if(delta < Integer.MIN_VALUE || delta > Integer.MAX_VALUE) {
continue;
}
int diff = (int) delta;
int prev = dp[j].getOrDefault(diff, 0);
int curr = dp[i].getOrDefault(diff, 0);
dp[i].put(diff, curr + prev + 1);
result += prev;
}
}
return result;
}
}
python:
class Solution:
def numberOfArithmeticSlices(self, A: List[int]) -> int:
dp = [{} for _ in range(len(A))]
result = 0
for i in range(len(A)):
for j in range(i):
delta = A[i] - A[j]
prev = dp[j].get(delta, 0)
curr = dp[i].get(delta, 0)
dp[i][delta]= curr + prev + 1
result += prev return result
Idea 3. 前面我们提到如果以3位数为base case这个并不好计算,换一个角度nums[i] - nums[j] = nums[j] - nums[k], 0 <= k < j < i, 如果有nums[k] = nums[j] * 2 - nums[i], 需要快速地找到nums[k],我们需要一个map记录nums[k] 和 index k.
dp[i][j] = sum(dp[j][k] + 1)
base case dp[i][j] = 0
result = sum(dp[i][j])
我们来走一下例子:[1, 1, 2, 3, 4, 5]
lookup(nums[k], [k]): 1-> [0, 1] , 2-> [2], 3-> [3], 4-> [4], 5-> [5]
i = 2, j = 1, nums[k] = 0, 不存在;
i = 3, j = 1, nums[k] = 2 * 1 - 3= -1,不存在; j= 2, nums[k] = 2 * 2 - 3 = 1, dp[3][2] += dp[2][0] + 1 + dp[2][1] + 1 = 2, sequence [1,2,3], [1, 2, 3]
i = 4, j = 1, nums[k] = 2 * 1 - 4 = -2, 不存在; j= 2, nums[k] = 2 * 2 - 4 = 0, 不存在; j = 3, nums[k] = 2 * 3 - 4 = 2, dp[4][3] += dp[3][2] + 1 = 3, sequence: [1,2,3, 4], [1, 2, 3, 4], [2, 3, 4]
i = 5, j = 1, nums[k] = 2 * 1 - 5 = -3, 不存在; j= 2, nums[k] = 2 * 2 - 5 = -1, 不存在; j = 3, nums[k] = 2 * 3 - 5 = 1, dp[5][3] = dp[3][1] + 1 + dp[3][0] + 1 = 2; j = 4, nums[k] = 2 * 4 - 5 = 3, dp[5][4] += dp[4][3] + 1 = 4, sequence: [1, 3, 5], [1, 3, 5], [1, 2, 3, 4, 5], [1, 2,3,4,5], [2, 3, 4, 5], [3, 4, 5]
Time complexity: O(n3) the worest case to loop the map lookup could be nearly as O(n), when have lots of duplicates like 1, 1, 1, 1, 2, 3, 4
Space complexity: O(n2)
class Solution {
public int numberOfArithmeticSlices(int[] A) {
int result = 0;
Map<Integer, List<Integer>> lookUp = new HashMap<>();
int[][] dp = new int[A.length][A.length];
for(int i = 0; i < A.length; ++i) {
if(lookUp.get(A[i]) == null) {
lookUp.put(A[i], new ArrayList<>());
}
lookUp.get(A[i]).add(i);
}
for(int i = 2; i < A.length; ++i) {
for(int j = 1; j < i; ++j) {
long tempTarget = 2 * (long)A[j] - A[i];
if(tempTarget < Integer.MIN_VALUE
|| tempTarget > Integer.MAX_VALUE) {
continue;
}
int target = (int) tempTarget;
if(lookUp.containsKey(target)) {
for(int k: lookUp.get(target)) {
if(k < j) {
dp[i][j] += dp[j][k] + 1;
}
}
result += dp[i][j];
}
}
}
return result;
}
}
python
class Solution:
def numberOfArithmeticSlices(self, A: List[int]) -> int:
result = 0
dp = [collections.defaultdict(int) for _ in range(len(A))]
lookup = collections.defaultdict(list) for i, val in enumerate(A):
lookup[val].append(i) for i in range(2, len(A)):
for j in range(1, i):
target = 2 * A[j] - A[i]
if target in lookup:
for k in lookup[target]:
if k < j:
dp[i][j] += dp[j][k] + 1 result += dp[i][j] return result
Arithmetic Slices II - Subsequence LT446的更多相关文章
- [LeetCode] Arithmetic Slices II - Subsequence 算数切片之二 - 子序列
A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...
- LeetCode 446. Arithmetic Slices II - Subsequence
原题链接在这里:https://leetcode.com/problems/arithmetic-slices-ii-subsequence/ 题目: A sequence of numbers is ...
- Leetcode: Arithmetic Slices II - Subsequence
A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...
- [Swift]LeetCode446. 等差数列划分 II - 子序列 | Arithmetic Slices II - Subsequence
A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...
- LeetCode446. Arithmetic Slices II - Subsequence
A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...
- 446. Arithmetic Slices II - Subsequence
A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...
- 446 Arithmetic Slices II - Subsequence 算数切片之二 - 子序列
详见:https://leetcode.com/problems/arithmetic-slices-ii-subsequence/description/ C++: class Solution { ...
- 第六周 Leetcode 446. Arithmetic Slices II - Subsequence (HARD)
Leetcode443 题意:给一个长度1000内的整数数列,求有多少个等差的子数列. 如 [2,4,6,8,10]有7个等差子数列. 想了一个O(n^2logn)的DP算法 DP[i][j]为 对于 ...
- [LeetCode] Arithmetic Slices 算数切片
A sequence of number is called arithmetic if it consists of at least three elements and if the diffe ...
随机推荐
- spring分页
1.Brand 商品品牌类 public class Brand { private Integer id; private String name; private String descripti ...
- 【C++】operator new/new operator/placement new之间的区别
new operator new operator即是c++中的关键字new.比如A* = new A; 中的new就是new operator. 它执行了三个步骤: 1. 分配内存空间 事实上,分配 ...
- Windows Server2012 R2 安装.NET Framework 3.5失败解决方法
转载:https://blog.csdn.net/F12138_/article/details/80220698 显示需要指定备用路径,但我没有指定 然后就出现了的失败T T! 由于我无法访问安装盘 ...
- 修改.net反编译的dll
用.Net reflector 打开,配合reflexil工具. 有两种修改方法. 1.重写,试过,但不好用. 2.修改IL指令 一般只需修改简单的if判断. 方法:找到需要修改的行,把brfalse ...
- Python之逻辑回归模型来预测
建立一个逻辑回归模型来预测一个学生是否被录取. import numpy as np import pandas as pd import matplotlib.pyplot as plt impor ...
- pta l1-49(天梯赛座位分配)
题目链接:https://pintia.cn/problem-sets/994805046380707840/problems/994805081289900032 题意:给定n个高校,每个高校m[i ...
- Python+Selenium学习--操作测试对象
场景 前面已经讲解了如果定位对象,定位之后需要对这个对象进行操作.是鼠标点击还是键盘输入,取决于我们定位的对象缩支持的操作. webdriver中比较常用的操作元素的方法有下面几个: clear ...
- 【转】关于easyui的窗口和tab页面不执行js说明
原地址:http://www.jeasyuicn.com/post-49.html 一直以来群里里面很多人反应,在用tab加载界面的时候,界面里面的js不会执行.今天GodSon在此说明一下原因. 不 ...
- Python: 调用youtube_dl实现视频下载
研究PySide与youtube_dl结合实现视频下载,抽丝剥蚕,步步维艰,却也颇有意思. 记录初始心得.界面以PySide之Qt编写,调用youtube_dl下载,回调出下载进度,代码如下: # e ...
- 【Linux 进程】孤儿进程、僵尸进程和守护进程
1.孤儿进程: 孤儿进程:一个父进程退出,而它的一个或多个子进程还在运行,那么那些子进程将成为孤儿进程.孤儿进程将被init进程(进程号为1)所收养,并由init进程对它们完成状态收集工作.孤儿进程是 ...