{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 一.算法流程\n",
    "adaboost回归模型与分类模型类似，主要的不同点在于错误率的计算、基模型的权重计算以及样本权重的更新，下面就直接介绍算法流程部分\n",
    "\n",
    "输入：训练集$T=\\{(x_1,y_1),(x_2,y_2),...,(x_N,y_N)\\}$，其中$x_i\\in R^n,y_i\\in R,i=1,2,...,N$   \n",
    "\n",
    "输出：最终回归模型$G(x)$  \n",
    "\n",
    ">（1）初始化训练数据的权重分布：  \n",
    "$$\n",
    "D_1=(w_{11},...,w_{1i},...,w_{1N}),w_{1i}=\\frac{1}{N},i=1,2,...,N\n",
    "$$  \n",
    "\n",
    ">（2）对$m=1,2,...,M:$   \n",
    "\n",
    ">>（2.1）使用具有权重分布$D_m$的训练数据集学习，得到基回归模型：$G_m(x)$   \n",
    "\n",
    ">>（2.2）计算$G_m(x)$在训练集上的误差率：  \n",
    "\n",
    ">>>(2.2.1)计算训练集上的最大误差：$E_m=max\\mid y_i-G_m(x_i)\\mid,i=1,2,...,N$  \n",
    "\n",
    ">>>(2.2.2)计算每个样本的相对误差，这里有三种计算方式可选：  \n",
    ">>>> a)线性误差：$e_{mi}=\\frac{\\mid y_i-G_m(x_i)\\mid}{E_m},i=1,2,...,N$  \n",
    "\n",
    ">>>> b)平方误差：$e_{mi}=\\frac{(y_i-G_m(x_i))^2}{E_m^2},i=1,2,...,N$  \n",
    "\n",
    ">>>> c)指数误差：$e_{mi}=1-exp(\\frac{-\\mid y_i-G_m(x_i)\\mid}{E_m}),i=1,2,...,N$\n",
    "\n",
    ">>>(2.2.3)计算误差率：$e_m=\\sum_{i=1}^N w_{mi}e_{mi},i=1,2,...,N$\n",
    "\n",
    ">>（2.3）计算$G_m(x)$的权重系数：$\\alpha_m=\\frac {e_m}{1-e_m}$  \n",
    "\n",
    ">>（2.4）更新训练样本权重：  \n",
    "$$\n",
    "w_{m+1,i}=\\frac{w_{mi}}{Z_m}\\alpha_m^{1-e_{mi}},i=1,2,...,N\n",
    "$$  \n",
    "这里$Z_m$是归一化因子   \n",
    "\n",
    ">（3）最终强学习器：  \n",
    "\n",
    "$$\n",
    "G(x)=\\sum_{i=1}^M \\frac {ln\\frac{1}{\\alpha_m}}{L}G_m(x),L=\\sum_{i=1}^M ln\\frac{1}{\\alpha_m}\n",
    "$$\n",
    "\n",
    "### 二.代码实现"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "os.chdir('../')\n",
    "from ml_models.tree import CARTRegressor\n",
    "import copy\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "%matplotlib inline\n",
    "\n",
    "\n",
    "class AdaBoostRegressor(object):\n",
    "    def __init__(self, base_estimator=None, n_estimators=10, learning_rate=1.0):\n",
    "        \"\"\"\n",
    "        :param base_estimator: 基学习器，允许异质；异质的情况下使用列表传入比如[estimator1,estimator2,...,estimator10],这时n_estimators会失效；\n",
    "                                同质的情况，单个estimator会被copy成n_estimators份\n",
    "        :param n_estimators: 基学习器迭代数量\n",
    "        :param learning_rate: 学习率，降低后续基学习器的权重，避免过拟合\n",
    "        \"\"\"\n",
    "        self.base_estimator = base_estimator\n",
    "        self.n_estimators = n_estimators\n",
    "        self.learning_rate = learning_rate\n",
    "        if self.base_estimator is None:\n",
    "            # 默认使用决策树桩\n",
    "            self.base_estimator = CARTRegressor(max_depth=2)\n",
    "        # 同质分类器\n",
    "        if type(base_estimator) != list:\n",
    "            estimator = self.base_estimator\n",
    "            self.base_estimator = [copy.deepcopy(estimator) for _ in range(0, self.n_estimators)]\n",
    "        # 异质分类器\n",
    "        else:\n",
    "            self.n_estimators = len(self.base_estimator)\n",
    "\n",
    "        # 记录estimator权重\n",
    "        self.estimator_weights = []\n",
    "\n",
    "        # 记录最终中位数值弱学习器的index\n",
    "        self.median_index = None\n",
    "\n",
    "    def fit(self, x, y):\n",
    "        n_sample = x.shape[0]\n",
    "        sample_weights = np.asarray([1.0] * n_sample)\n",
    "        for index in range(0, self.n_estimators):\n",
    "            self.base_estimator[index].fit(x, y, sample_weight=sample_weights)\n",
    "\n",
    "            errors = np.abs(self.base_estimator[index].predict(x) - y)\n",
    "            error_max = np.max(errors)\n",
    "\n",
    "            # 计算线性误差，其他误差类型，可以自行扩展\n",
    "            linear_errors = errors / error_max\n",
    "            # 计算误分率\n",
    "            error_rate = np.dot(linear_errors, sample_weights / n_sample)\n",
    "\n",
    "            # 计算权重系数\n",
    "            alpha_rate = error_rate / (1.0 - error_rate + 1e-10)\n",
    "            self.estimator_weights.append(alpha_rate)\n",
    "\n",
    "            # 更新样本权重\n",
    "            for j in range(0, n_sample):\n",
    "                sample_weights[j] = sample_weights[j] * np.power(alpha_rate, 1 - linear_errors[j])\n",
    "            sample_weights = sample_weights / np.sum(sample_weights) * n_sample\n",
    "\n",
    "        # 更新estimator权重\n",
    "        self.estimator_weights = np.log(1 / np.asarray(self.estimator_weights))\n",
    "        for i in range(0, self.n_estimators):\n",
    "            self.estimator_weights[i] *= np.power(self.learning_rate, i)\n",
    "        self.estimator_weights /= np.sum(self.estimator_weights)\n",
    "\n",
    "    def predict(self, x):\n",
    "        return np.sum(\n",
    "            [self.estimator_weights[i] * self.base_estimator[i].predict(x) for i in\n",
    "             range(0, self.n_estimators)],\n",
    "            axis=0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "#构造数据\n",
    "data = np.linspace(1, 10, num=100)\n",
    "target = np.sin(data) + np.random.random(size=100)#添加噪声\n",
    "data = data.reshape((-1, 1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x1ae127d3198>]"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x1ae036df780>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#训练模型\n",
    "model=AdaBoostRegressor(base_estimator=CARTRegressor(max_bins=20),n_estimators=10)\n",
    "model.fit(data,target)\n",
    "plt.scatter(data, target)\n",
    "plt.plot(data, model.predict(data), color='r')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}