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
- Java中递归的优缺点,Java写一个递归遍历目录下面的所有文件包括子文件夹里边的文件。
题目: 遍历出aaa文件夹下的文件 首先分析思路: 1.首先判断这个文件夹是否为文件,通过isFile()函数可以判断是否为文件. 2.然后通过isDirectory判断是否为目录. 3.如果是目录就 ...
- SMB扫描
server message block协议,Windows特有的一个协议,实现较复杂,windows应用最广的一个协议,也是安全问题最多的问题,smb协议windows默认开发,用于文件共享. sm ...
- 8 种 NoSQL 数据库系统对比(转自: http://blog.jobbole.com/1344/)
导读:Kristóf Kovács 是一位软件架构师和咨询顾问,他最近发布了一片对比各种类型NoSQL数据库的文章. 虽然SQL数据库是非常有用的工具,但经历了15年的一支独秀之后垄断即将被打破.这只 ...
- NumPy 字符串函数
NumPy 字符串函数 以下函数用于对 dtype 为 numpy.string_ 或 numpy.unicode_ 的数组执行向量化字符串操作. 它们基于 Python 内置库中的标准字符串函数. ...
- 198. House Robber(Array; DP)
You are a professional robber planning to rob houses along a street. Each house has a certain amount ...
- SQL Server中row_number的用法
ROW_NUMBER()函数将针对SELECT语句返回的每一行,从1开始编号,赋予其连续的编号.在查询时应用了一个排序标准后,只有通过编号才能够保证其顺序是一致的,当使用ROW_NUMBER函数时,也 ...
- 安装程序遇到错误0x80240037
安装ie插件或者微软的一些其他程序等报错0x80240037,如下图所示: 解决办法: 下面我将以ie的插件IE11-Windows6.1-KB3008923-x64.msu安装为例: 1.把IE11 ...
- 微信小程序——过滤器的模拟
>> 编写wxs文件——filter.wxs //1. 价格格式化 function getPriceFormat(value) { return parseFloat(isNaN(val ...
- [Java学习]常用类-包装类型
八种基本类型对应的包装类 Java中的数据类型由八种基本类型,以及引用类型组成. byte short int long float double boolbean char Object 为了方便, ...
- 成对使用new和delete,传值传引用
首先: delete []p;是用来删除对象数组的,特别是你声明的是对象数组!!! 如果new中用了[],delete一定要用[]:在new中没有使用,在delete中一定不要使用. 其次: 当你使用 ...