1 files changed, 156 insertions, 0 deletions
diff --git a/perceptron/perceptron.cc b/perceptron/perceptron.cc
new file mode 100644
index 0000000..3aa6a72
--- /dev/null
+++ b/perceptron/perceptron.cc
@@ -0,0 +1,156 @@
+#include <iostream>
+#include <vector>
+#include <utility>
+#include <string>
+#include <boost/assign/std/vector.hpp>
+#include <cassert>
+#include <cmath>
+
+using namespace std;
+using namespace boost::assign;
+
+typedef vector<double> vec;
+
+
+/*
+ * features
+ *
+ */
+double
+isupper_f(string s)
+{
+  if ( isupper(s[0]) ) return 0;
+  return 1;
+}
+
+double
+len_f(string s)
+{
+  return s.size();
+}
+
+
+/*
+ * vector ops
+ *
+ */
+double
+dotprod( vec a, vec b )
+{
+  assert( a.size() == b.size() );
+  double sum;
+  for ( size_t i = 0; i < a.size(); i++ ) {
+    sum += a[i] * b[i];
+  }
+  return sum;
+}
+
+vec
+vadd( double x, vec v )
+{
+  for ( size_t i = 0; i < v.size(); i++ ) {
+    v[i] += x;
+  }
+  return v;
+}
+
+vec
+vadd( vec v, vec w )
+{
+  assert( v.size() == w.size() );
+  for ( size_t i = 0; i < v.size(); i++ ) {
+    v[i] += w[i];
+  }
+  return v;
+}
+
+vec
+vsub( vec v, vec w )
+{
+  assert(v.size() == w.size());
+  for (size_t i = 0; i < v.size(); i++) {
+    v[i] -= w[i];
+  }
+  return v;
+}
+
+vec
+vmult( double x, vec v )
+{
+  for (size_t i = 0; i < v.size(); i++) {
+    v[i] *= x;
+  }
+  return v;
+}
+
+double
+vlen( vec v )
+{
+  return sqrt( dotprod(v, v) );
+}
+
+vec
+vnorm( vec v )
+{
+  double len = vlen(v);
+  for (size_t i = 0; i < v.size(); i++) {
+    v[i] /= len;
+  }
+  return v;
+}
+
+
+int
+main( void )
+{
+  vec classes;
+  vector<string> words;
+  classes +=  -1.,   1.,     -1.,    1.,     -1.,   1.,     -1.,   1.,      -1.,     1.;
+  words   += "Haus","house","Mauer","plant","Zeug","hello","Maus","worker","Quartz","point";
+
+  vector<pair<double, string> > examples;
+  for (size_t i = 0; i < classes.size(); i++) {
+    pair<double, string> p;
+    p.first = classes[i];
+    p.second = words[i];
+    examples.push_back(p);
+  }
+
+  vec weights;
+  weights += 0.,0.,0.;
+
+  double eta = 0.5;
+  
+  for (size_t t = 0; t < 2; t++)
+  {
+    for (size_t x = 0; x < examples.size(); x++) {
+      vec feats;
+      feats += 1.;
+      feats += isupper_f(examples[x].second), len_f(examples[x].second);
+      double err = dotprod(weights, feats);
+      //if ( (examples[x].first == 1 && err < 0) || (examples[x].first == -1 && err >= 0) ) {
+      if ( examples[x].first * err <= 0 ) {
+        weights = vadd(weights, vmult(examples[x].first*eta, feats));
+      }
+      cout << "w " << weights[0] << " " << weights[1] << " " << weights[2] << endl;
+    }
+    cout << "--" << endl;
+  }
+  
+   cout << endl;
+   for (size_t x = 0; x < examples.size(); x++) {
+      vec feats;
+      feats += 1.;
+      feats += isupper_f(examples[x].second), len_f(examples[x].second);
+      cout << dotprod(weights, feats) << endl;
+    }
+    
+    vec f;
+    f += 1.;
+    f += isupper_f("Bauer"), len_f("Bauer");
+    cout << "test: " << dotprod(weights, f) << endl;
+
+
+  return 0;
+}
+