Update lda.ipynb

Author: csuldw

LDA/lda.ipynb

@@ -105,28 +105,28 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "[[18.93314346 18.34477196]\n",
-      " [16.30593804 15.62170289]\n",
-      " [17.10822678 16.34611882]\n",
-      " [18.4907721  15.80400898]\n",
-      " [18.14483149 15.85428922]\n",
-      " [16.97320081 17.2396569 ]\n",
-      " [18.73810721 15.11751718]\n",
-      " [16.4681786  19.3626708 ]] [[12.80631524  9.36789595]\n",
-      " [11.5139152   8.69570105]\n",
-      " [ 8.28460266  8.98015663]\n",
-      " [ 8.55723255  8.30610269]\n",
-      " [10.63855077  8.75788812]\n",
-      " [10.9176947  12.91157592]\n",
-      " [10.10859137  8.15180198]\n",
-      " [ 8.50446587  9.13270925]]\n"
+      "[[16.96464962 16.94502501]\n",
+      " [15.05454667 17.09013667]\n",
+      " [15.56489963 19.84229387]\n",
+      " [16.43831074 15.90474953]\n",
+      " [18.48476088 15.03913508]\n",
+      " [15.82917037 15.11968731]\n",
+      " [15.73431849 19.21742217]\n",
+      " [19.2702835  19.73339472]] [[2.63292147 3.47120986]\n",
+      " [6.76104203 5.98453328]\n",
+      " [6.62851068 6.8916269 ]\n",
+      " [3.77655065 3.40796527]\n",
+      " [2.3063607  3.40134092]\n",
+      " [3.09106307 3.41137181]\n",
+      " [3.29697416 2.28073533]\n",
+      " [4.97203486 3.18686582]]\n"
      ]
     }
    ],
    "source": [
     "def createDataSet():  \n",
     "    X1 = np.mat(np.random.random((8, 2)) * 5 + 15)  #类别A\n",
-    "    X2 = np.mat(np.random.random((8, 2)) * 5 + 8)   #类别B\n",
+    "    X2 = np.mat(np.random.random((8, 2)) * 5 + 2)   #类别B\n",
     "    return X1, X2  \n",
     "\n",
     "x1, x2 = createDataSet()\n",
@@ -150,24 +150,24 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "dataX: [[18.93314346 18.34477196]\n",
-      " [16.30593804 15.62170289]\n",
-      " [17.10822678 16.34611882]\n",
-      " [18.4907721  15.80400898]\n",
-      " [18.14483149 15.85428922]\n",
-      " [16.97320081 17.2396569 ]\n",
-      " [18.73810721 15.11751718]\n",
-      " [16.4681786  19.3626708 ]\n",
-      " [12.80631524  9.36789595]\n",
-      " [11.5139152   8.69570105]\n",
-      " [ 8.28460266  8.98015663]\n",
-      " [ 8.55723255  8.30610269]\n",
-      " [10.63855077  8.75788812]\n",
-      " [10.9176947  12.91157592]\n",
-      " [10.10859137  8.15180198]\n",
-      " [ 8.50446587  9.13270925]]\n",
-      "w: [[0.77542443]\n",
-      " [0.63144039]]\n"
+      "dataX: [[16.96464962 16.94502501]\n",
+      " [15.05454667 17.09013667]\n",
+      " [15.56489963 19.84229387]\n",
+      " [16.43831074 15.90474953]\n",
+      " [18.48476088 15.03913508]\n",
+      " [15.82917037 15.11968731]\n",
+      " [15.73431849 19.21742217]\n",
+      " [19.2702835  19.73339472]\n",
+      " [ 2.63292147  3.47120986]\n",
+      " [ 6.76104203  5.98453328]\n",
+      " [ 6.62851068  6.8916269 ]\n",
+      " [ 3.77655065  3.40796527]\n",
+      " [ 2.3063607   3.40134092]\n",
+      " [ 3.09106307  3.41137181]\n",
+      " [ 3.29697416  2.28073533]\n",
+      " [ 4.97203486  3.18686582]]\n",
+      "w: [[0.76741174]\n",
+      " [0.6411546 ]]\n"
      ]
     }
    ],
@@ -199,9 +199,9 @@
    "outputs": [
     {
      "data": {
-      "image/png": "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\n",
+      "image/png": "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\n",
       "text/plain": [
-       "<matplotlib.figure.Figure at 0x9121a58>"
+       "<matplotlib.figure.Figure at 0x90d9a90>"
       ]
      },
      "metadata": {},
@@ -229,22 +229,22 @@
      "name": "stdout",
      "output_type": "stream",
      "text": [
-      "Output:  [[25.69014016 25.72919165]]\n"
+      "Output:  [[-2.40824965]]\n"
      ]
     }
    ],
    "source": [
-    "test2=np.mat([22,22])  \n",
-    "g=np.dot(w.T, test2.T - 0.5*(meanX(x1)-meanX(x2)))  \n",
+    "test2 = np.mat([2, 8]) \n",
+    "g = np.dot(w.T, test2.T - 0.5 * (meanX(x1)-meanX(x2)).T)  \n",
     "print \"Output: \", g  "
    ]
   },
   {
-   "cell_type": "code",
-   "execution_count": null,
+   "cell_type": "markdown",
    "metadata": {},
-   "outputs": [],
-   "source": []
+   "source": [
+    "结果输出表示该样本为负样本。"
+   ]
   }
  ],
  "metadata": {