446. Arithmetic Slices II - Subsequence
Hard

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

  1. [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 ...

  2. 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 ...

  3. Leetcode: Arithmetic Slices II - Subsequence

    A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...

  4. [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 ...

  5. LeetCode446. Arithmetic Slices II - Subsequence

    A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...

  6. 446. Arithmetic Slices II - Subsequence

    A sequence of numbers is called arithmetic if it consists of at least three elements and if the diff ...

  7. 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 { ...

  8. 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 ...

  9. [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

  1. git库上传

    1.第一步,拉下项目 2.添加 3.提交到本地和仓库

  2. webservice客户端 get delete post 请求

    package com.cn.eport.util.common; import java.io.IOException; import java.util.List; import org.apac ...

  3. Aptana在Eclipse的安装

    1.下载 com.aptana.rcp.product-3.4.2.zip文件 https://pan.baidu.com/s/1sl81Vit 2.安装 接着Next.直到成功 3.怎么判定安装成功 ...

  4. java面试题:jvm

    jvm内存区域 Q:jvm内存怎么划分的? 答: 方法区(线程共享):各个线程共享的一个区域,用于存储虚拟机加载的类信息.常量.静态变量.即时编译器编译后的代码等数据.虽然 Java 虚拟机规范把方法 ...

  5. NIO简单理解

    NIO:新IO,同步的非阻塞IO. 1.Java NIO 由以下几个核心部分组成:Channels(通道).Buffers(缓冲区).Selectors(选择器) Channels(通道) 1.所有的 ...

  6. Android 性能测试之内存 --- 追加腾讯性能案例,安卓抓取性能扫盲帖

    内存测试: 思路 目前做的是酒店APP,另下载安装几个个第三方酒店的APP以方便对比(相当于可以做竞品测试) 数据的获取来源是ADB底层命令,而且最好是不需要root权限,因为很多手机root很麻烦或 ...

  7. HDU 1698 Just a Hook(线段树区间更新查询)

    描述 In the game of DotA, Pudge’s meat hook is actually the most horrible thing for most of the heroes ...

  8. [剑指Offer]42-连续子数组的最大和(DP)

    题目链接 https://www.nowcoder.com/practice/459bd355da1549fa8a49e350bf3df484?tpId=13&tqId=11183&t ...

  9. mysql学习1:数据类型:数字型,日期和时间,字符串类型(总结)

    mysql数据类型:数字型,日期和时间,字符串类型 摘要 MySQL中定义数据字段的类型对数据库的优化是非常重要的: MySQL支持多种类型,大致可以分为三类,如下. 数字类型 整数:tinyint. ...

  10. RabbitMQ 参数们的Power “续”

      参数中的 arguments 之前讲参数的一些作用的时候,忽略了最后一个字典类型的参数,因为这个参数是大有文章的,值得单独进出来说道说道. 这时,就不得不打开我们的 Web UI管理系统了,可以看 ...