2014-04-29 00:20
题目:给定一个长字符串,和一个词典。如果允许你将长串分割成若干个片段,可能会存在某些片段在词典里查不到,有些则查得到。请设计算法进行分词,使得查不到的片段个数最少。
解法:用空间换取时间的动态规划算法,首先用O(n^2)的时间判断每一个片段是否在字典里。这个过程其实可以通过字典树来进行加速,时间上能优化一个阶,不过我没写,偷懒用<unordered_set>代表了字典。之后通过O(n)时间的动态规划,dp[i]表示当前位置的查不到的片段的最少个数。对于懂代码的人,代码说的比文字清楚,所以请看代码。
代码:
1 // 17.14 Given a dictionary of words, and a long string. You may find a way to cut the string into words, where some of them may or may not be in the dictionary. 2 // Dynamic programming is a good thing, but trades space in for time. 3 #include <iostream> 4 #include <string> 5 #include <unordered_set> 6 #include <vector> 7 using namespace std; 8 9 int main() 10 { 11 string data; 12 unordered_set<string> dict; 13 vector<vector<bool> > contains; 14 vector<int> dp; 15 int i, j; 16 string s; 17 int n; 18 int tmp; 19 20 while (cin >> data && data != "") { 21 cin >> n; 22 for (i = 0; i < n; ++i) { 23 cin >> s; 24 dict.insert(s); 25 } 26 n = (int)data.length(); 27 28 contains.resize(n); 29 for (i = 0; i < n; ++i) { 30 contains[i].resize(n); 31 } 32 for (i = 0; i < n; ++i) { 33 s = ""; 34 for (j = i; j < n; ++j) { 35 s.push_back(data[j]); 36 contains[i][j] = (dict.find(s) != dict.end()); 37 } 38 } 39 40 dp.resize(n); 41 for (i = 0; i < n; ++i) { 42 dp[i] = contains[0][i] ? 0 : i + 1; 43 for (j = 0; j < i; ++j) { 44 tmp = dp[j] + (contains[j + 1][i] ? 0 : i - j); 45 dp[i] = dp[i] < tmp ? dp[i] : tmp; 46 } 47 } 48 49 printf("%d\n", dp[n - 1]); 50 51 for (i = 0; i < n; ++i) { 52 contains[i].clear(); 53 } 54 contains.clear(); 55 dp.clear(); 56 dict.clear(); 57 } 58 59 return 0; 60 }