解决方案C-ABU量化系统.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ABU 量化系统  简介（版本 0.1）\n",
    "\n",
    "\n",
    "* #### __author__ = 'BBFamily'\n",
    "* ####  __email__ = 'bbfamily@126.com'\n",
    "* ####  __weixin__ = 'aaaabbbuu'"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 第九部分 解决方案C"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ‘非均衡胜负收益’带来的必然’非均衡胜负比例‘，目标由’因子‘的能力解决一部分，’模式识别‘提升关键的一部分"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import ZEnv\n",
    "import ZLog\n",
    "import ZCommonUtil\n",
    "import pandas as pd\n",
    "pd.options.display.max_columns = 100\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "fn = './data/cache/orders_pd_ump_hit_predict_abu'\n",
    "key = 'orders_pd_ump_hit_predict_abu'\n",
    "orders_pd_ump = ZCommonUtil.load_hdf5(fn, key)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "这里的orders_pd_ump数据由 Factor中的各个裁决组成pipeline具体请查询代码\n",
    "\n",
    "\n",
    "            g_pp_dict = {\n",
    "            UmpMainClass(None, MlFiterMainPdClass, predict=True): {'w_col': MlFiterMainPd.g_w_col, 'need_ind_cnt': 1},\n",
    "            UmpMainClass(None, MlFiterWavePdClass, predict=True): {'w_col': MlFiterWavePd.g_w_col, 'need_ind_cnt': 1},\n",
    "            UmpMainClass(None, MlFiterDegPdClass, predict=True): {'w_col': MlFiterDegPd.g_w_col, 'need_ind_cnt': 1},\n",
    "\n",
    "            UmpJumpClass(None, MlFiterJumpPdClass, predict=True): {'w_col': MlFiterJumpPd.g_w_col, 'need_ind_cnt': 1,\n",
    "                                                                   'cons': lambda order: order['diff_days'] < 21},\n",
    "\n",
    "            UmpEdgeClass(None, predict=True): {}\n",
    "            }\n",
    "            \n",
    "            g_pipe_line = UmpPipeLineClass(g_pp_dict)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**UmpEdge 边裁**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "import UmpEdge"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "ump_edge = UmpEdge.UmpEdgeClass(orders_pd_ump)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**边裁使用profit， profit_cg作为gmm分类数据生成ss分类序列，之后根据profit_cg进行rank数据生成p_rk_cg，在找到top winN，top loss\n",
    "N N现在的设置是25%且对外不暴露，分别给于1， -1, 其它的都是0生成rk列，将atr，deg，wave所有数据的mumpy矩阵保存起来，对输入的数据进行标准化后实行距离对比，找到最匹配的\n",
    "rk标签**\n",
    "\n",
    "\n",
    "1. 优化距离对比方法现在是pairwise_distances，相似度，协整等方式\n",
    "2. 类似knn，找k个相思点投票，不止先只有一个"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>profit</th>\n",
       "      <th>profit_cg</th>\n",
       "      <th>atr_std</th>\n",
       "      <th>deg_hisWindowPd</th>\n",
       "      <th>deg_windowPd</th>\n",
       "      <th>deg_60WindowPd</th>\n",
       "      <th>wave_score1</th>\n",
       "      <th>wave_score2</th>\n",
       "      <th>wave_score3</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>21051.90</td>\n",
       "      <td>0.105290</td>\n",
       "      <td>0.045208</td>\n",
       "      <td>-1.256691</td>\n",
       "      <td>3.895622</td>\n",
       "      <td>5.363046</td>\n",
       "      <td>0.116619</td>\n",
       "      <td>-0.122372</td>\n",
       "      <td>-0.222840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>-8411.46</td>\n",
       "      <td>-0.043313</td>\n",
       "      <td>0.234763</td>\n",
       "      <td>15.908454</td>\n",
       "      <td>4.108007</td>\n",
       "      <td>4.199374</td>\n",
       "      <td>0.646118</td>\n",
       "      <td>0.731843</td>\n",
       "      <td>0.662114</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>-19158.06</td>\n",
       "      <td>-0.095806</td>\n",
       "      <td>-0.408504</td>\n",
       "      <td>21.175689</td>\n",
       "      <td>3.394158</td>\n",
       "      <td>7.384808</td>\n",
       "      <td>0.050669</td>\n",
       "      <td>-0.131306</td>\n",
       "      <td>-0.158877</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>-5845.79</td>\n",
       "      <td>-0.074492</td>\n",
       "      <td>0.687850</td>\n",
       "      <td>-7.023154</td>\n",
       "      <td>-0.075254</td>\n",
       "      <td>5.413456</td>\n",
       "      <td>0.671606</td>\n",
       "      <td>0.799534</td>\n",
       "      <td>0.829494</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>12345.00</td>\n",
       "      <td>0.061728</td>\n",
       "      <td>0.011969</td>\n",
       "      <td>11.519025</td>\n",
       "      <td>3.621430</td>\n",
       "      <td>5.765467</td>\n",
       "      <td>-0.149591</td>\n",
       "      <td>-0.227755</td>\n",
       "      <td>-0.259943</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              profit  profit_cg   atr_std  deg_hisWindowPd  deg_windowPd  \\\n",
       "2015-07-28  21051.90   0.105290  0.045208        -1.256691      3.895622   \n",
       "2015-07-28  -8411.46  -0.043313  0.234763        15.908454      4.108007   \n",
       "2015-07-28 -19158.06  -0.095806 -0.408504        21.175689      3.394158   \n",
       "2015-07-28  -5845.79  -0.074492  0.687850        -7.023154     -0.075254   \n",
       "2015-07-28  12345.00   0.061728  0.011969        11.519025      3.621430   \n",
       "\n",
       "            deg_60WindowPd  wave_score1  wave_score2  wave_score3  \n",
       "2015-07-28        5.363046     0.116619    -0.122372    -0.222840  \n",
       "2015-07-28        4.199374     0.646118     0.731843     0.662114  \n",
       "2015-07-28        7.384808     0.050669    -0.131306    -0.158877  \n",
       "2015-07-28        5.413456     0.671606     0.799534     0.829494  \n",
       "2015-07-28        5.765467    -0.149591    -0.227755    -0.259943  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ump_edge.fiter.df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "rk        -1         0    1\n",
      "ss                         \n",
      "0   1.000000  0.000000  0.0\n",
      "6   1.000000  0.000000  0.0\n",
      "7   0.892651  0.107349  0.0\n",
      "9   1.000000  0.000000  0.0\n",
      "19  1.000000  0.000000  0.0\n",
      "rk   -1    0    1\n",
      "ss               \n",
      "2   0.0  0.0  1.0\n",
      "4   0.0  0.0  1.0\n",
      "5   0.0  0.0  1.0\n",
      "10  0.0  0.0  1.0\n",
      "11  0.0  0.0  1.0\n",
      "13  0.0  0.0  1.0\n",
      "16  0.0  0.0  1.0\n",
      "18  0.0  0.0  1.0\n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA64AAAH7CAYAAADM2sIfAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAABJRU5ErkJggg==\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x119d42b10>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "xt, xt_pct = ump_edge.gmm_component_filter()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>rk</th>\n",
       "      <th>-1</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ss</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2465</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>4475</td>\n",
       "      <td>92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>639</td>\n",
       "      <td>4472</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>3053</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>47</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>3243</td>\n",
       "      <td>390</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>37</td>\n",
       "      <td>4865</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>886</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>978</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0</td>\n",
       "      <td>1775</td>\n",
       "      <td>2239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0</td>\n",
       "      <td>3679</td>\n",
       "      <td>606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0</td>\n",
       "      <td>4347</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>213</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>1541</td>\n",
       "      <td>2887</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1723</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>104</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "rk    -1     0     1\n",
       "ss                  \n",
       "0   2465     0     0\n",
       "1      0  4475    92\n",
       "2      0     0     4\n",
       "3    639  4472     0\n",
       "4      0     0  3053\n",
       "5      0     0    33\n",
       "6     47     0     0\n",
       "7   3243   390     0\n",
       "8     37  4865     0\n",
       "9    886     0     0\n",
       "10     0     0    20\n",
       "11     0     0   978\n",
       "12     0  1775  2239\n",
       "13     0     0     2\n",
       "14     0  3679   606\n",
       "15     0  4347     0\n",
       "16     0     0   213\n",
       "17  1541  2887     0\n",
       "18     0     0  1723\n",
       "19   104     0     0"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xt"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th>rk</th>\n",
       "      <th>-1</th>\n",
       "      <th>0</th>\n",
       "      <th>1</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>ss</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.979855</td>\n",
       "      <td>0.020145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.125024</td>\n",
       "      <td>0.874976</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.892651</td>\n",
       "      <td>0.107349</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.007548</td>\n",
       "      <td>0.992452</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.442202</td>\n",
       "      <td>0.557798</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.858576</td>\n",
       "      <td>0.141424</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>0.348013</td>\n",
       "      <td>0.651987</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "rk        -1         0         1\n",
       "ss                              \n",
       "0   1.000000  0.000000  0.000000\n",
       "1   0.000000  0.979855  0.020145\n",
       "2   0.000000  0.000000  1.000000\n",
       "3   0.125024  0.874976  0.000000\n",
       "4   0.000000  0.000000  1.000000\n",
       "5   0.000000  0.000000  1.000000\n",
       "6   1.000000  0.000000  0.000000\n",
       "7   0.892651  0.107349  0.000000\n",
       "8   0.007548  0.992452  0.000000\n",
       "9   1.000000  0.000000  0.000000\n",
       "10  0.000000  0.000000  1.000000\n",
       "11  0.000000  0.000000  1.000000\n",
       "12  0.000000  0.442202  0.557798\n",
       "13  0.000000  0.000000  1.000000\n",
       "14  0.000000  0.858576  0.141424\n",
       "15  0.000000  1.000000  0.000000\n",
       "16  0.000000  0.000000  1.000000\n",
       "17  0.348013  0.651987  0.000000\n",
       "18  0.000000  0.000000  1.000000\n",
       "19  1.000000  0.000000  0.000000"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "xt_pct"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "ump_edge.dump_clf()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**利用测试裁决数据，寻找最优裁决参数**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import ast\n",
    "def map_str_dict(extra_info, key):\n",
    "    try:\n",
    "        map_ast = ast.literal_eval(extra_info)[key]\n",
    "    except Exception, e:\n",
    "        import pdb\n",
    "        pdb.set_trace()\n",
    "        raise e\n",
    "    return map_ast\n",
    "\n",
    "def extra_info_to_pd(orders_pd):\n",
    "    orders_pd['ump_main_mlfitermainpdclass_predict'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfitermainpdclass_predict',))\n",
    "    orders_pd['ump_main_mlfiterdegpdclass_predict'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfiterdegpdclass_predict',))\n",
    "    orders_pd['ump_main_mlfiterwavepdclass_predict'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfiterwavepdclass_predict',))\n",
    "    orders_pd['ump_jump_mlfiterjumppdclass_predict'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_jump_mlfiterjumppdclass_predict',))\n",
    "\n",
    "    orders_pd['ump_main_mlfitermainpdclass_hit'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfitermainpdclass_hit',))\n",
    "    orders_pd['ump_main_mlfiterdegpdclass_hit'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfiterdegpdclass_hit',))\n",
    "    orders_pd['ump_main_mlfiterwavepdclass_hit'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_main_mlfiterwavepdclass_hit',))\n",
    "    orders_pd['ump_jump_mlfiterjumppdclass_hit'] = orders_pd['ExtraInfo'].apply(map_str_dict, args=('./data/cache/ump_jump_mlfiterjumppdclass_hit',))\n",
    "\n",
    "# extra_info_to_pd(orders_pd_ump)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "(47374, 40)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>buy Date</th>\n",
       "      <th>buy Price</th>\n",
       "      <th>buy Cnt</th>\n",
       "      <th>buyFactor</th>\n",
       "      <th>Sell Date</th>\n",
       "      <th>Sell Price</th>\n",
       "      <th>Sell Type</th>\n",
       "      <th>Symbol</th>\n",
       "      <th>MaxLoss</th>\n",
       "      <th>ExtraInfo</th>\n",
       "      <th>key</th>\n",
       "      <th>profit</th>\n",
       "      <th>result</th>\n",
       "      <th>sell_type_extra</th>\n",
       "      <th>R</th>\n",
       "      <th>sharpe</th>\n",
       "      <th>profit_cg</th>\n",
       "      <th>profit_cg_hunder</th>\n",
       "      <th>keep_days</th>\n",
       "      <th>lowBkCnt</th>\n",
       "      <th>his_miss_days</th>\n",
       "      <th>atr_std</th>\n",
       "      <th>deg_hisWindowPd</th>\n",
       "      <th>deg_windowPd</th>\n",
       "      <th>deg_60WindowPd</th>\n",
       "      <th>jump_power</th>\n",
       "      <th>diff_days</th>\n",
       "      <th>wave_score1</th>\n",
       "      <th>wave_score2</th>\n",
       "      <th>wave_score3</th>\n",
       "      <th>symbol_index</th>\n",
       "      <th>ump_main_mlfitermainpdclass_predict</th>\n",
       "      <th>ump_main_mlfiterdegpdclass_predict</th>\n",
       "      <th>ump_main_mlfiterwavepdclass_predict</th>\n",
       "      <th>ump_jump_mlfiterjumppdclass_predict</th>\n",
       "      <th>ump_main_mlfitermainpdclass_hit</th>\n",
       "      <th>ump_main_mlfiterdegpdclass_hit</th>\n",
       "      <th>ump_main_mlfiterwavepdclass_hit</th>\n",
       "      <th>ump_jump_mlfiterjumppdclass_hit</th>\n",
       "      <th>ind_key</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2012-07-12</th>\n",
       "      <td>20120712</td>\n",
       "      <td>22.465</td>\n",
       "      <td>8902</td>\n",
       "      <td>BuyGoldenFactorClass:42</td>\n",
       "      <td>20120720</td>\n",
       "      <td>23.75</td>\n",
       "      <td>win</td>\n",
       "      <td>usPFE</td>\n",
       "      <td>20.90</td>\n",
       "      <td>{'./data/cache/ump_jump_mlfiterjumppdclass_pre...</td>\n",
       "      <td>497</td>\n",
       "      <td>11439.07</td>\n",
       "      <td>1</td>\n",
       "      <td>BuyGoldenFactorClass:42: winPrice :above950</td>\n",
       "      <td>0.821086</td>\n",
       "      <td>3.8237</td>\n",
       "      <td>0.05720</td>\n",
       "      <td>5.720009</td>\n",
       "      <td>8</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.373831</td>\n",
       "      <td>12.933343</td>\n",
       "      <td>1.104452</td>\n",
       "      <td>-0.163254</td>\n",
       "      <td>-1.096786</td>\n",
       "      <td>16</td>\n",
       "      <td>0.027985</td>\n",
       "      <td>-0.042228</td>\n",
       "      <td>-0.140315</td>\n",
       "      <td>6</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>47372</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2012-07-16</th>\n",
       "      <td>20120716</td>\n",
       "      <td>3.155</td>\n",
       "      <td>63391</td>\n",
       "      <td>BuyGoldenFactorClass:42</td>\n",
       "      <td>20120719</td>\n",
       "      <td>2.58</td>\n",
       "      <td>loss</td>\n",
       "      <td>usBAMM</td>\n",
       "      <td>2.25</td>\n",
       "      <td>{'./data/cache/ump_jump_mlfiterjumppdclass_pre...</td>\n",
       "      <td>499</td>\n",
       "      <td>-36449.82</td>\n",
       "      <td>-1</td>\n",
       "      <td>mv_pre_atr preClose 2.0</td>\n",
       "      <td>-0.635359</td>\n",
       "      <td>-1.67571</td>\n",
       "      <td>-0.18225</td>\n",
       "      <td>-18.225037</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0.025823</td>\n",
       "      <td>8.891965</td>\n",
       "      <td>0.686205</td>\n",
       "      <td>7.232708</td>\n",
       "      <td>-1.260642</td>\n",
       "      <td>73</td>\n",
       "      <td>-0.213653</td>\n",
       "      <td>-0.317415</td>\n",
       "      <td>-0.258648</td>\n",
       "      <td>6</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>47373</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            buy Date  buy Price  buy Cnt                buyFactor Sell Date  \\\n",
       "2012-07-12  20120712     22.465     8902  BuyGoldenFactorClass:42  20120720   \n",
       "2012-07-16  20120716      3.155    63391  BuyGoldenFactorClass:42  20120719   \n",
       "\n",
       "            Sell Price Sell Type  Symbol  MaxLoss  \\\n",
       "2012-07-12       23.75       win   usPFE    20.90   \n",
       "2012-07-16        2.58      loss  usBAMM     2.25   \n",
       "\n",
       "                                                    ExtraInfo  key    profit  \\\n",
       "2012-07-12  {'./data/cache/ump_jump_mlfiterjumppdclass_pre...  497  11439.07   \n",
       "2012-07-16  {'./data/cache/ump_jump_mlfiterjumppdclass_pre...  499 -36449.82   \n",
       "\n",
       "            result                              sell_type_extra         R  \\\n",
       "2012-07-12       1  BuyGoldenFactorClass:42: winPrice :above950  0.821086   \n",
       "2012-07-16      -1                      mv_pre_atr preClose 2.0 -0.635359   \n",
       "\n",
       "             sharpe  profit_cg  profit_cg_hunder  keep_days  lowBkCnt  \\\n",
       "2012-07-12   3.8237    0.05720          5.720009          8         1   \n",
       "2012-07-16 -1.67571   -0.18225        -18.225037          3         1   \n",
       "\n",
       "            his_miss_days   atr_std  deg_hisWindowPd  deg_windowPd  \\\n",
       "2012-07-12              0  0.373831        12.933343      1.104452   \n",
       "2012-07-16              0  0.025823         8.891965      0.686205   \n",
       "\n",
       "            deg_60WindowPd  jump_power  diff_days  wave_score1  wave_score2  \\\n",
       "2012-07-12       -0.163254   -1.096786         16     0.027985    -0.042228   \n",
       "2012-07-16        7.232708   -1.260642         73    -0.213653    -0.317415   \n",
       "\n",
       "            wave_score3  symbol_index ump_main_mlfitermainpdclass_predict  \\\n",
       "2012-07-12    -0.140315             6                                True   \n",
       "2012-07-16    -0.258648             6                                True   \n",
       "\n",
       "           ump_main_mlfiterdegpdclass_predict  \\\n",
       "2012-07-12                               True   \n",
       "2012-07-16                               True   \n",
       "\n",
       "           ump_main_mlfiterwavepdclass_predict  \\\n",
       "2012-07-12                                True   \n",
       "2012-07-16                                True   \n",
       "\n",
       "           ump_jump_mlfiterjumppdclass_predict  \\\n",
       "2012-07-12                                True   \n",
       "2012-07-16                                True   \n",
       "\n",
       "            ump_main_mlfitermainpdclass_hit  ump_main_mlfiterdegpdclass_hit  \\\n",
       "2012-07-12                                0                               0   \n",
       "2012-07-16                                0                               0   \n",
       "\n",
       "            ump_main_mlfiterwavepdclass_hit  ump_jump_mlfiterjumppdclass_hit  \\\n",
       "2012-07-12                                0                                0   \n",
       "2012-07-16                                0                                0   \n",
       "\n",
       "            ind_key  \n",
       "2012-07-12    47372  \n",
       "2012-07-16    47373  "
      ]
     },
     "execution_count": 107,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ZLog.info(orders_pd_ump.shape)\n",
    "orders_pd_ump.tail(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "from MlFiter import MlFiterClass\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**寻找主裁最优参数**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>result</th>\n",
       "      <th>ump_main_mlfitermainpdclass_predict</th>\n",
       "      <th>ump_main_mlfiterdegpdclass_predict</th>\n",
       "      <th>ump_main_mlfiterwavepdclass_predict</th>\n",
       "      <th>ump_main_mlfitermainpdclass_hit</th>\n",
       "      <th>ump_main_mlfiterdegpdclass_hit</th>\n",
       "      <th>ump_main_mlfiterwavepdclass_hit</th>\n",
       "      <th>ind_key</th>\n",
       "      <th>predict_sum</th>\n",
       "      <th>hit_sum</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>7</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>523</td>\n",
       "      <td>2</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>25</td>\n",
       "      <td>0</td>\n",
       "      <td>524</td>\n",
       "      <td>2</td>\n",
       "      <td>50</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>540</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>547</td>\n",
       "      <td>2</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2015-07-28</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>549</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            result  ump_main_mlfitermainpdclass_predict  \\\n",
       "2015-07-28       0                                    0   \n",
       "2015-07-28       0                                    1   \n",
       "2015-07-28       0                                    1   \n",
       "2015-07-28       0                                    0   \n",
       "2015-07-28       0                                    0   \n",
       "\n",
       "            ump_main_mlfiterdegpdclass_predict  \\\n",
       "2015-07-28                                   1   \n",
       "2015-07-28                                   0   \n",
       "2015-07-28                                   1   \n",
       "2015-07-28                                   1   \n",
       "2015-07-28                                   1   \n",
       "\n",
       "            ump_main_mlfiterwavepdclass_predict  \\\n",
       "2015-07-28                                    1   \n",
       "2015-07-28                                    1   \n",
       "2015-07-28                                    0   \n",
       "2015-07-28                                    1   \n",
       "2015-07-28                                    1   \n",
       "\n",
       "            ump_main_mlfitermainpdclass_hit  ump_main_mlfiterdegpdclass_hit  \\\n",
       "2015-07-28                                7                               0   \n",
       "2015-07-28                                0                              25   \n",
       "2015-07-28                                0                               0   \n",
       "2015-07-28                                3                               0   \n",
       "2015-07-28                                1                               0   \n",
       "\n",
       "            ump_main_mlfiterwavepdclass_hit  ind_key  predict_sum  hit_sum  \n",
       "2015-07-28                                0      523            2        7  \n",
       "2015-07-28                                0      524            2       50  \n",
       "2015-07-28                                1      540            2        0  \n",
       "2015-07-28                                0      547            2        3  \n",
       "2015-07-28                                0      549            2        1  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orders_pd_ump['ind_key'] = np.arange(0, len(orders_pd_ump))\n",
    "orders_pd_tmp = orders_pd_ump.filter(regex='result|ump_main_mlfiter*|ind_key')\n",
    "order_has_ret = orders_pd_tmp[orders_pd_tmp['result'] <> 0]\n",
    "order_has_ret['result'] = np.where(order_has_ret['result'] == -1, 0, 1)\n",
    "\n",
    "order_has_ret['ump_main_mlfitermainpdclass_predict'] = np.where(order_has_ret['ump_main_mlfitermainpdclass_predict'] == True, 1, 0)\n",
    "order_has_ret['ump_main_mlfiterdegpdclass_predict'] = np.where(order_has_ret['ump_main_mlfiterdegpdclass_predict'] == True, 1, 0)\n",
    "order_has_ret['ump_main_mlfiterwavepdclass_predict'] = np.where(order_has_ret['ump_main_mlfiterwavepdclass_predict'] == True, 1, 0)\n",
    "# order_has_ret['ump_main_mlfiterjumppdclass_predict'] = np.where(order_has_ret['ump_main_mlfiterjumppdclass_predict'] == True, 1, 0)\n",
    "\n",
    "order_has_ret = order_has_ret[(order_has_ret['ump_main_mlfitermainpdclass_predict'] == 0) | \n",
    "                              (order_has_ret['ump_main_mlfiterdegpdclass_predict'] == 0) | \n",
    "                              (order_has_ret['ump_main_mlfiterwavepdclass_predict'] == 0)]\n",
    "\n",
    "order_has_ret['predict_sum'] = order_has_ret['ump_main_mlfitermainpdclass_predict'] + order_has_ret['ump_main_mlfiterdegpdclass_predict'] + \\\n",
    "            order_has_ret['ump_main_mlfiterwavepdclass_predict']\n",
    "\n",
    "order_has_ret['hit_sum'] = order_has_ret['ump_main_mlfiterdegpdclass_hit'] + order_has_ret['ump_main_mlfitermainpdclass_hit'] + \\\n",
    "            order_has_ret['ump_main_mlfiterdegpdclass_hit']\n",
    "    \n",
    "matrix = order_has_ret.as_matrix()\n",
    "y = matrix[:, 0]\n",
    "x = matrix[:, 1:]\n",
    "fiter = MlFiterClass(x, y, order_has_ret)\n",
    "fiter.df.head()\n",
    "order_has_ret.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "accuracy mean: 0.636323529412\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "array([ 0.6375    ,  0.6375    ,  0.63865546,  0.63865546,  0.62605042,\n",
       "        0.63445378,  0.63445378,  0.63865546,  0.63865546,  0.63865546])"
      ]
     },
     "execution_count": 111,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fiter.estimator.svc()\n",
    "# fiter.estimator.random_forest_classifier()\n",
    "# fiter.estimator.bagging_classifier()\n",
    "fiter.cross_val_accuracy_score()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**1个hit以上就上0.65 20个hit 0.70以上，> 20个就可以直接使用裁决了，小于的等待，辅助裁判，边裁**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.colorbar.Colorbar at 0x10c910350>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAX0AAAEACAYAAABfxaZOAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n6EVi0A8hhF7k/0bZusu+7CSsAAAAAElFTkSuQmCC\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11abc7c90>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "import matplotlib.pyplot as plt\n",
    "\n",
    "hd_range = np.arange(0, 50)\n",
    "hd_result = []\n",
    "hd_sum = []\n",
    "for hd in hd_range:\n",
    "    # xt = order_has_ret[(order_has_ret['hit_sum'] > hd) & (order_has_ret['predict_sum'] > 1)]['result'].value_counts()\n",
    "    xt = order_has_ret[(order_has_ret['hit_sum'] > hd)]['result'].value_counts()\n",
    "    hs = xt.sum()\n",
    "    hd_sum.append(hs)\n",
    "    hd_result.append(float(xt[0])/hs)\n",
    "    \n",
    "cmap = plt.get_cmap('jet', 20)\n",
    "cmap.set_under('gray')\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(hd_range, hd_result)\n",
    "cax = ax.scatter(hd_range, hd_result, c=hd_sum, cmap=cmap, vmin=np.min(hd_sum),\n",
    "                 vmax=np.max(hd_sum))\n",
    "ax.grid(True)\n",
    "fig.colorbar(cax, label='hd_sum', extend='min')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**本身数据就少，辅助裁判尽量裁决 jump ump可以使用hit 5作为阀值，大于5个hit直接裁决，否则等待边裁**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "[ 0  1  2  3  4  7  8 15 18]\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.colorbar.Colorbar at 0x11bf25510>"
      ]
     },
     "execution_count": 119,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAXAAAAEACAYAAACqOy3+AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nOVehvIA751yF8gLunHMVygu4c85VqP8PQNHeV9Q/GP8AAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11b9f7c50>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "orders_pd_tmp = orders_pd_ump.filter(regex='result|ump_jump_mlfiter|ind_key*')\n",
    "order_has_ret = orders_pd_tmp[orders_pd_tmp['result'] <> 0]\n",
    "order_has_ret['result'] = np.where(order_has_ret['result'] == -1, 0, 1)\n",
    "order_has_ret['ump_jump_mlfiterjumppdclass_predict'] = np.where(order_has_ret['ump_jump_mlfiterjumppdclass_predict'] == True, 1, 0)\n",
    "\n",
    "order_has_ret = order_has_ret[(order_has_ret['ump_jump_mlfiterjumppdclass_predict'] == 0)]\n",
    "\n",
    "hd_range = np.unique(order_has_ret['ump_jump_mlfiterjumppdclass_hit'])[:-1]\n",
    "# -1 要使用0开始的范围\n",
    "hd_range = np.array(hd_range) - 1\n",
    "print hd_range\n",
    "hd_result = []\n",
    "hd_sum = []\n",
    "for hd in hd_range:\n",
    "    xt = order_has_ret[order_has_ret['ump_jump_mlfiterjumppdclass_hit'] > hd]['result'].value_counts()\n",
    "    hs = xt.sum()\n",
    "    hd_sum.append(hs)\n",
    "    hd_result.append(float(xt[0])/hs)\n",
    "    \n",
    "cmap = plt.get_cmap('jet', 20)\n",
    "cmap.set_under('gray')\n",
    "fig, ax = plt.subplots()\n",
    "ax.plot(hd_range, hd_result)\n",
    "cax = ax.scatter(hd_range, hd_result, c=hd_sum, cmap=cmap, vmin=np.min(hd_sum),\n",
    "                 vmax=np.max(hd_sum))\n",
    "ax.grid(True)\n",
    "fig.colorbar(cax, label='hd_sum', extend='min')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "jump_hit_pd = order_has_ret[order_has_ret['ump_jump_mlfiterjumppdclass_hit'] < 3]\n",
    "jump_hit_pd_org = orders_pd_ump.iloc[jump_hit_pd['ind_key']]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def apply_judge(order):\n",
    "    return ump_edge.predict(atr_std=order.atr_std, deg_hisWindowPd=order.deg_hisWindowPd, deg_windowPd=order.deg_windowPd,\n",
    "                 deg_60WindowPd=order.deg_60WindowPd, wave_score1=order.wave_score1, wave_score2=order.wave_score2,\n",
    "                 wave_score3=order.wave_score3)\n",
    "jump_hit_pd['edge'] = jump_hit_pd_org.apply(apply_judge, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       " 0    175\n",
       "-1     12\n",
       " 1      7\n",
       "Name: edge, dtype: int64"
      ]
     },
     "execution_count": 122,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "jump_hit_pd['edge'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**这里边裁的裁决只是为了验证正确性，由于边裁predict使用的是输入于原始数据矩阵的度量判断， 如果对裁决范围数量不满意，可以修改\n",
    "edge top loss，top win范围**\n",
    "\n",
    "        con_x = np.concatenate((x, dump_clf['fiter_x']), axis=0)\n",
    "        x_scale_param = self.scaler.fit(con_x)\n",
    "        con_x = self.scaler.fit_transform(con_x, x_scale_param)\n",
    "        distance_min_ind = pairwise_distances(con_x[0].reshape(1, -1), con_x[1:],\n",
    "                                              metric='euclidean').argmin()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 123,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from sklearn import metrics\n",
    "jump_hit_pd['result2'] = np.where(jump_hit_pd['result'] == 0, -1, 1)\n",
    "metrics.accuracy_score(jump_hit_pd[jump_hit_pd['edge'] == -1]['result2'],\n",
    "                                           jump_hit_pd[jump_hit_pd['edge'] == -1]['edge'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 124,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.accuracy_score(jump_hit_pd[jump_hit_pd['edge'] == 1]['result2'],\n",
    "                                           jump_hit_pd[jump_hit_pd['edge'] == 1]['edge'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2166, 40)"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_hit_pd = order_has_ret[order_has_ret['hit_sum'] < 20]\n",
    "main_hit_pd_org = orders_pd_ump.iloc[main_hit_pd['ind_key']]\n",
    "main_hit_pd_org.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "collapsed": false
   },
   "outputs": [],
   "source": [
    "main_hit_pd['edge'] = main_hit_pd_org.apply(apply_judge, axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       " 0    1298\n",
       "-1     557\n",
       " 1     311\n",
       "Name: edge, dtype: int64"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_hit_pd['edge'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.0"
      ]
     },
     "execution_count": 132,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "main_hit_pd['result2'] = np.where(main_hit_pd['result'] == 0, -1, 1)\n",
    "metrics.accuracy_score(main_hit_pd[main_hit_pd['edge'] == -1]['result2'],\n",
    "                                           main_hit_pd[main_hit_pd['edge'] == -1]['edge'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "collapsed": false
   },
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "all fit order = (44906, 40)\n",
      "win rate = 0.500757137131\n",
      "profit_cg.sum() = 272.117613217\n",
      "win mean = 0.0743788075658 loss_mean = -0.0626291433739 \n"
     ]
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYoAAAELCAYAAADHksFtAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x11a86dfd0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "from UmpMain import UmpMainClass\n",
    "from MlFiterGoldenPd import MlFiterGoldenPdClass\n",
    "UmpMainClass(orders_pd_ump, MlFiterGoldenPdClass).show_general()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "anaconda-cloud": {},
  "kernelspec": {
   "display_name": "Python [Root]",
   "language": "python",
   "name": "Python [Root]"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 0
}