Created
June 9, 2013 01:52
-
-
Save spaghetti-source/5737292 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| // n-grams graph G = (V,E) | |
| // V: set of n-grams (i.e., n-tuple of words for n = 1 .. N ) | |
| // E: u->v iff v is a next n-gram of v in the document | |
| // | |
| // Usage: ./a.out < pg11.txt | |
| // ( http://www.gutenberg.org/cache/epub/11/pg11.txt ) | |
| // | |
| #include <iostream> | |
| #include <vector> | |
| #include <cstdio> | |
| #include <cstdlib> | |
| #include <map> | |
| #include <cmath> | |
| #include <cstring> | |
| #include <numeric> | |
| #include <functional> | |
| #include <algorithm> | |
| #include <tr1/unordered_map> | |
| using namespace std; | |
| using namespace tr1; | |
| #define ALL(c) c.begin(), c.end() | |
| #define FOR(i,c) for(typeof(c.begin())i=c.begin();i!=c.end();++i) | |
| #define REP(i,n) for(int i=0;i<n;++i) | |
| #define fst first | |
| #define snd second | |
| typedef string NGram; // separated by _ (underscore) | |
| unordered_map<string, int> wordToId; | |
| vector<string> idToWord; | |
| vector< vector<int> > adj; | |
| void addWord(string s) { | |
| if (wordToId.count(s)) return; | |
| int n = idToWord.size(); | |
| wordToId[s] = n; | |
| idToWord.push_back(s); | |
| adj.push_back( vector<int>() ); | |
| } | |
| void addEdge(string s, string t) { | |
| addWord(s); | |
| addWord(t); | |
| adj[ wordToId[s] ].push_back( wordToId[t] ); | |
| } | |
| void addSentense(vector<string> sentense) { | |
| int nmax = 4; | |
| int l = sentense.size(); | |
| // REP(i, l) cout << sentense[i] << " "; cout << endl; | |
| for (int p = 0; p+1 < l; ++p) { // word1 word2 | |
| vector<string> prev; | |
| prev.push_back(sentense[p]); | |
| for (int i = 1; i < nmax && p-i >= 0; ++i) | |
| prev.push_back( sentense[p-i] + " " + prev.back() ); | |
| string next = sentense[p+1]; | |
| FOR(s, prev) addEdge(*s, next); | |
| for (int i = 1; i < nmax && p+1+i < l; ++i) { | |
| next = next + " " + sentense[p+1+i]; | |
| FOR(s, prev) addEdge(*s, next); | |
| //FOR(s, prev) REP(k, i+1) addEdge(*s, next); | |
| } | |
| } | |
| } | |
| void readFile(FILE *fp, int n) { | |
| vector<string> sentense; | |
| string s; | |
| for (char c; (c = getchar()) != EOF; ) { | |
| if (isalpha(c)) { | |
| s += tolower(c); | |
| } else if (s != "") { | |
| sentense.push_back(s); | |
| s = ""; | |
| if (c == '.' || c == ',' || c == '?' || c == '!' || c == ':') { | |
| addSentense(sentense); | |
| sentense.clear(); | |
| } | |
| } | |
| } | |
| } | |
| void pagerank(vector<double> b = vector<double>(adj.size(), 1), double a = 0.85) { | |
| vector<double> x = b; | |
| REP(epoch, 100) { | |
| vector<double> y(x.size()); | |
| REP(u, adj.size()) { | |
| int d = adj[u].size(); | |
| REP(k, d) { | |
| int v = adj[u][k]; | |
| y[v] += x[u] / d; | |
| } | |
| } | |
| double err = 0; | |
| REP(u, x.size()) { | |
| double rhs = a*y[u] + (1-a)*b[u]; | |
| err += pow(x[u]-rhs, 2); | |
| x[u] = rhs; | |
| } | |
| err = sqrt(err); | |
| if (err < 1e-6) break; | |
| } | |
| double total = 0; | |
| REP(u, x.size()) total += x[u]; | |
| REP(u, x.size()) x[u] /= total; | |
| vector< pair<double, int> > rank; | |
| REP(u, x.size()) rank.push_back( make_pair(x[u], u) ); | |
| sort(ALL(rank), greater< pair<double,int> >()); | |
| REP(k, min(rank.size(), 200u)) { | |
| string word = idToWord[rank[k].snd]; | |
| double score = rank[k].fst; | |
| cout << word << " " << score << endl; | |
| } | |
| } | |
| int main() { | |
| readFile(stdin, 3); | |
| vector<double> b(adj.size()); | |
| b[ wordToId["cat"] ] = 1; | |
| pagerank(b); | |
| //pagerank(); | |
| } |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment