{ "cells": [ { "cell_type": "code", "execution_count": 1, "metadata": { "collapsed": true }, "outputs": [], "source": [ "# %load /Users/facai/Study/book_notes/preconfig.py\n", "%matplotlib inline\n", "\n", "import matplotlib.pyplot as plt\n", "import seaborn as sns\n", "sns.set(color_codes=True)\n", "sns.set(font='SimHei')\n", "plt.rcParams['axes.grid'] = False\n", "\n", "from IPython.display import SVG\n", "\n", "def show_image(filename, figsize=None):\n", " if figsize:\n", " plt.figure(figsize=figsize)\n", "\n", " plt.imshow(plt.imread(filename))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "评价函数模块 `_criterion.*` 详解\n", "=================================" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 0. 大纲\n", "\n", "本文会按照类的继承关系,由上到下介绍各个类的成员及实现细节,尽量说得详细点。" ] }, { "cell_type": "code", "execution_count": 17, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/svg+xml": [ "Criterion+proxy_impurity_improvement()+impurity_improvement()ClassificationCriterion+init()+reset()+reverse_reset()+update()+node_value()RegressionCriterion+init()+reset()+reverse_reset()+update()+node_value()Entropy+node_impurity()+children_impurity()Gini+node_impurity()+children_impurity()MSE+node_impurity()+children_impurity()+proxy_impurity_improvement()FriedmanMSE+proxy_impurity_improvement()+impurity_improvement()" ], "text/plain": [ "" ] }, "execution_count": 17, "metadata": {}, "output_type": "execute_result" } ], "source": [ "SVG(\"./res/uml/Model___criterion_1.svg\")" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 1. Criterion, 接口类\n", "Criterion 作为接口之用,相当于 Java 中的抽象类。其接口定义在 `_criterion.pxd` 文件,实现在 `_criterion.pyx` 中。\n", "\n", "我们对 Criterion 分为两部份来说明:内部成员,和函数方法。\n", "\n", "##### 1.0 内部成员\n", "为了计算和存储效率,sklearn 做了些针对性的数据设计。当然,其实思路挺直觉的,我们带着具体的成员变量来说明。\n", "\n", "+ 数据分条 (stride)\n", "\n", " 二维数组会用一维数组(其实是个连续的缓存)加上偏移量来实现,我以前在数字图像处理中常看到。 \n", " \n", " ```Python\n", " 26 # Internal structures\n", " 27 cdef DOUBLE_t* y # Values of y\n", " 28 cdef SIZE_t y_stride # Stride in y (since n_outputs >= 1)\n", " 29 cdef DOUBLE_t* sample_weight # Sample weights\n", " ```\n", "\n", " 分类的标签或者回归的值($y$)可能是个向量,所以需要对 $y$ 值分条,每条的长度自然是 `y_stride`。 \n", " `sample_weight` 是各个样本的权重值,不用多讲。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "+ 使用索引\n", "\n", " 决策树分割节点时会涉及到移动样本,因为 $x$ 通常维数较多,开销较大。对于这类问题,一个常见的做法就是交换索引,而不变动值。具体到数组操组,这个索引就是下标值。\n", " \n", " ```Python\n", " 31 cdef SIZE_t* samples # Sample indices in X, y\n", " 32 cdef SIZE_t start # samples[start:pos] are the samples in the left node\n", " 33 cdef SIZE_t pos # samples[pos:end] are the samples in the right node\n", " 34 cdef SIZE_t end\n", " ```\n", " \n", " 这里样本值的索引区间其实是前闭后开,即 samples[start:pos) 和 samples[pos:end)。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "+ 加权处理\n", "\n", " 加权的样本数目统计\n", " ```Python\n", " 36 cdef SIZE_t n_outputs # Number of outputs\n", " 37 cdef SIZE_t n_node_samples # Number of samples in the node (end-start)\n", " 38 cdef double weighted_n_samples # Weighted number of samples (in total)\n", " 39 cdef double weighted_n_node_samples # Weighted number of samples in the node\n", " 40 cdef double weighted_n_left # Weighted number of samples in the left node\n", " 41 cdef double weighted_n_right # Weighted number of samples in the right node\n", " ```\n", " \n", " 加权的 $y$ 值累加和(若是分类,则是对各个类别单独做累加)\n", " ```Python\n", " 43 cdef double* sum_total # For classification criteria, the sum of the\n", " 44 # weighted count of each label. For regression,\n", " 45 # the sum of w*y. sum_total[k] is equal to\n", " 46 # sum_{i=start}^{end-1} w[samples[i]]*y[samples[i], k],\n", " 47 # where k is output index.\n", " 48 cdef double* sum_left # Same as above, but for the left side of the split\n", " 49 cdef double* sum_right # same as above, but for the right side of the split\n", " ```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 1.1 函数方法\n", "+ 纯接口\n", "\n", "```Python\n", " 54 # Methods\n", " 55 cdef void init(self, DOUBLE_t* y, SIZE_t y_stride, DOUBLE_t* sample_weight,\n", " 56 double weighted_n_samples, SIZE_t* samples, SIZE_t start,\n", " 57 SIZE_t end) nogil\n", " 58 cdef void reset(self) nogil\n", " 59 cdef void reverse_reset(self) nogil\n", " 60 cdef void update(self, SIZE_t new_pos) nogil\n", " 61 cdef double node_impurity(self) nogil\n", " 62 cdef void children_impurity(self, double* impurity_left,\n", " 63 double* impurity_right) nogil\n", " 64 cdef void node_value(self, double* dest) nogil\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "+ 计算不纯度提升量\n", "\n", "```Python\n", " 65 cdef double impurity_improvement(self, double impurity) nogil\n", " 66 cdef double proxy_impurity_improvement(self) nogil\n", "```\n", "\n", "`proxy_impurity_improvement`:\n", "\n", "\\begin{align}\n", " & \\frac{N_t}{N} \\left ( \\text{ impurity } - \\frac{N_{t,R}}{N_t} \\text{ right_impurity } - \\frac{N_{t,L}}{N_t} \\text{ left_impurity } \\right ) \\\\\n", " &= \\frac{N_t}{N} \\text{ impurity } - \\frac{N_{t,R}}{N} \\text{ right_impurity } - \\frac{N_{t,L}}{N} \\text{ left_impurity } \\\\\n", " &= \\frac{N_t}{N} \\text{ impurity } - \\frac{1}{N} \\underbrace{ \\left ( N_{t,R} \\text{ right_impurity } - N_{t,L} \\text{ left_impurity } \\right ) }_{\\text{impurity_improvement}} \\\\\n", "\\end{align}\n", "\n", "对一个节点,去除公式中的常量,弄了个 `impurity_improvement` 来减少计算量。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 2. 分类评价函数" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### 2.0 ClassificationCriterion\n", "ClassificationCriterion 类是分类评估函数的父类,主要𫐄实现初始化和一些预处理函数。我们从上到下依序看代码。\n", "\n", "##### 2.0.0 变量\n", "\n", "```Python\n", " 207 cdef class ClassificationCriterion(Criterion):\n", " 208 \"\"\"Abstract criterion for classification.\"\"\"\n", " 209\n", " 210 cdef SIZE_t* n_classes\n", " 211 cdef SIZE_t sum_stride\n", "```\n", "\n", "这里分类为了支持多输出(target)问题,用 `n_classes` 来统计每个输出的类别个数,用 `sum_stride` 来对齐不同输出的结果。用文字表述比较困难,下面的图说明更清晰点。" ] }, { "cell_type": "code", "execution_count": 18, "metadata": { "collapsed": false }, "outputs": [ { "data": { "image/png": "iVBORw0KGgoAAAANSUhEUgAAAeIAAAI6CAYAAAD7bEt8AAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nwzAMwwQQFmKGYRiGCSAsxAzDMAwTQP4PKX7upOpZS9cAAAAASUVORK5CYII=\n", "text/plain": [ "" ] }, "metadata": {}, "output_type": "display_data" } ], "source": [ "show_image(\"./res/ClassificationCriterion.jpg\", figsize=(8,15))" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.1 `__cinit__`\n", "接下来,我们看构造函数:\n", "\n", "```Python\n", " 213 def __cinit__(self, SIZE_t n_outputs,\n", " 214 np.ndarray[SIZE_t, ndim=1] n_classes):\n", " 215 \"\"\"Initialize attributes for this criterion.\n", " 216 +-- 7 lines: Parameters---------------------------------------------------------\n", " 223 \"\"\"\n", " 224\n", " 225 #+-- 20 lines: self.y = NULL------------------------------------------------------\n", " 245\n", " 246 safe_realloc(&self.n_classes, n_outputs)\n", " 247 #+-- 6 lines: cdef SIZE_t k = 0--------------------------------------------------\n", " 253 for k in range(n_outputs):\n", " 254 self.n_classes[k] = n_classes[k]\n", " 255\n", " 256 if n_classes[k] > sum_stride:\n", " 257 sum_stride = n_classes[k]\n", " 258\n", " 259 self.sum_stride = sum_stride\n", " 260\n", " 261 cdef SIZE_t n_elements = n_outputs * sum_stride\n", " 262 self.sum_total = calloc(n_elements, sizeof(double))\n", " 263 #+-- 7 lines: self.sum_left = calloc(n_elements, sizeof(double))-------\n", "``` \n", "\n", "主要是三个工作:\n", "\n", "+ 初始化变量,设置默认值。\n", "+ 复制 `n_classes`,同时记录最大值 `sum_stride` 用于对齐。\n", "+ 为 `sum_*` 统计各类别信息的变量申请空间。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.2 `init`\n", "应该是为了重复使用,这里单独写了初始化函数。\n", "\n", "```Python\n", " 282 cdef void init(self, DOUBLE_t* y, SIZE_t y_stride,\n", " 283 DOUBLE_t* sample_weight, double weighted_n_samples,\n", " 284 SIZE_t* samples, SIZE_t start, SIZE_t end) nogil:\n", " 285 \"\"\"Initialize the criterion at node samples[start:end] and\n", " 286 +-- 19 lines: children samples[start:start] and samples[start:end].--------------\n", " 305 \"\"\"\n", " 306\n", " 307 #+-- 19 lines: self.y = y---------------------------------------------------------\n", " 326\n", " 327 for k in range(self.n_outputs):\n", " 328 memset(sum_total + offset, 0, n_classes[k] * sizeof(double))\n", " 329 offset += self.sum_stride\n", " 330\n", " 331 for p in range(start, end):\n", " 332 i = samples[p]\n", " 333 #+-- 4 lines: w is originally set to be 1.0, meaning that if no sample weights---\n", " 337 w = sample_weight[i]\n", " 338\n", " 339 # Count weighted class frequency for each target\n", " 340 for k in range(self.n_outputs):\n", " 341 c = y[i * y_stride + k]\n", " 342 sum_total[k * self.sum_stride + c] += w\n", " 343\n", " 344 self.weighted_n_node_samples += w\n", " 345\n", " 346 # Reset to pos=start\n", " 347 self.reset()\n", "```\n", "\n", "主要流程是:\n", "\n", "+ 307L-326L, 复制参数和设置默认值;\n", "+ 327L-330L, 清零 `sum_total`。 \n", " 注意 328L 的优化, `memset` 最后一个参数是各输出的实际类别数,所以只会清实际用到的空间。\n", "+ 331L-344L, 统计各类别的权重值 `sum_total` 和总权重值 `weighted_n_node_samples`。 \n", " 其中,340L 取到输出 $k$, 341L 取到类别 $c$,累加相应的总值 $\\text{sum_total}[k, c]$。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.3 `reset`\n", "reset 方法正如其名,做了重置工作,将节点位置 `pos` 恢复到最左边 `start` 位置,并更新统计信息。\n", "\n", "```Python\n", " 349 cdef void reset(self) nogil: 350 \"\"\"Reset the criterion at pos=start.\"\"\"\n", " 351\n", " 352 self.pos = self.start\n", " 353\n", " 354 self.weighted_n_left = 0.0\n", " 355 self.weighted_n_right = self.weighted_n_node_samples\n", " 356\n", " 357 #+-- 7 lines: cdef double* sum_total = self.sum_total----------------------------\n", " 364 for k in range(self.n_outputs):\n", " 365 memset(sum_left, 0, n_classes[k] * sizeof(double))\n", " 366 memcpy(sum_right, sum_total, n_classes[k] * sizeof(double))\n", " 367\n", " 368 sum_total += self.sum_stride\n", " 369 #+--- 2 lines: sum_left += self.sum_stride---------------------------------------\n", "```\n", "\n", "主要工作,清掉左节点的信息,覆盖右节点。\n", "\n", "`reverse_reset` 和 `reset` 操作互反,不赘述。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.4 `update`\n", "update 函数将节点移到指定位置,并更新统计信息。\n", "\n", "```Python\n", " 394 cdef void update(self, SIZE_t new_pos) nogil:\n", " 395 \"\"\"Updated statistics by moving samples[pos:new_pos] to the left child.\n", " 396 +-- 6 lines: Parameters---------------------------------------------------------\n", " 402 \"\"\"\n", " 403 cdef DOUBLE_t* y = self.y\n", " 404 #+-- 17 lines: cdef SIZE_t pos = self.pos-----------------------------------------\n", " 421\n", " 422 #+-- 8 lines: Update statistics up to new_pos------------------------------------\n", " 430 if (new_pos - pos) <= (end - new_pos):\n", " 431 for p in range(pos, new_pos):\n", " 432 #+-- 8 lines: i = samples[p]-----------------------------------------------------\n", " 440 sum_left[label_index] += w\n", " 441\n", " 442 self.weighted_n_left += w\n", " 443\n", " 444 else:\n", " 445 self.reverse_reset()\n", " 446\n", " 447 for p in range(end - 1, new_pos - 1, -1):\n", " 448 #+-- 8 lines: i = samples[p]-----------------------------------------------------\n", " 456 sum_left[label_index] -= w\n", " 457\n", " 458 self.weighted_n_left -= w\n", " 459\n", " 460 # Update right part statistics\n", " 461 self.weighted_n_right = self.weighted_n_node_samples - self.weighted_n_le ft\n", " 462 for k in range(self.n_outputs):\n", " 463 for c in range(n_classes[k]):\n", " 464 sum_right[c] = sum_total[c] - sum_left[c]\n", " 465 #+-- 4 lines: sum_right += self.sum_stride---------------------------------------\n", " 469\n", " 470 self.pos = new_pos \n", "```\n", "\n", "主要流程如下:\n", "\n", "+ 初始化参数;\n", "+ 422L-458L, 更新左节点的统计信息。 \n", " 这里的优化,是判断指定位置、当前节点和最右边三者的距离,遍历较小的距离。\n", "+ 460L-465L, 更新右节点的统计信息:$\\text{sum_right } = \\text{ sum_total } - \\text{ sum_left }$。 " ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.5 `node_value`\n", "作用是将 `sum_total` 的数据拷贝出来,不多说。\n", "\n", "```Python\n", " 479 cdef void node_value(self, double* dest) nogil:\n", " 480 \"\"\"Compute the node value of samples[start:end] and save it into dest.\n", " 481 +-- 5 lines: Parameters---------------------------------------------------------\n", " 486 \"\"\"\n", " 487 #+-- 4 lines: cdef double* sum_total = self.sum_total----------------------------\n", " 491\n", " 492 for k in range(self.n_outputs):\n", " 493 memcpy(dest, sum_total, n_classes[k] * sizeof(double))\n", " 494 #+-- 2 lines: dest += self.sum_stride--------------------------------------------\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 2.0.6 小结\n", "我们可以看出,`ClassificationCriterion` 的主要工作是在更新 `sum_*` 变量。这三个变量的数据是相应节点中各个类的权重值,用这三个变量可以算出各类 $k$ 在节点 $i$ 的权重占比 $p_i^k$。而我们知道,常用的 gini、entropy 和 misclassification_rate 都是基于 $p_i^k$ 来表达。所以,将更新权重信息函数汇总,便于复用,正是此父类的存在意义。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "有了 $p_k^i$ 权重占比,再下面的子类就纯粹是数学计算了,加上这几个式子都简单,代码非常清晰。需要注意,对多输出的问题,评价值是统一加合取平均。\n", "\n", "#### 2.1 Entropy\n", "\n", " $\\text{cross-entropy } = -\\sum_{k=0}^{K-1} \\text{ count_k } \\log( \\text{ count_k }) $\n", "\n", "```Python\n", " 498 cdef class Entropy(ClassificationCriterion):\n", " 499 \"\"\"Cross Entropy impurity criterion.\n", " 500 +-- 11 lines: This handles cases where the target is a classification taking valu\n", " 511 cross-entropy = -\\sum_{k=0}^{K-1} count_k log(count_k)\n", " 512 \"\"\"\n", " 513\n", " 514 cdef double node_impurity(self) nogil:\n", " 515 #+-- 10 lines: \"\"\"Evaluate the impurity of the current node, i.e. the impurity of-\n", " 525 for k in range(self.n_outputs):\n", " 526 for c in range(n_classes[k]):\n", " 527 count_k = sum_total[c]\n", " 528 if count_k > 0.0:\n", " 529 count_k /= self.weighted_n_node_samples\n", " 530 entropy -= count_k * log(count_k)\n", " 531 #+-- 3 lines: sum_total += self.sum_stride---------------------------------------\n", " 534 return entropy / self.n_outputs\n", " 535\n", " 536 cdef void children_impurity(self, double* impurity_left,\n", " 537 double* impurity_right) nogil:\n", " 538 #+-- 22 lines: \"\"\"Evaluate the impurity in children nodes-------------------------\n", " 560 for k in range(self.n_outputs):\n", " 561 for c in range(n_classes[k]):\n", " 562 count_k = sum_left[c]\n", " 563 if count_k > 0.0:\n", " 564 count_k /= self.weighted_n_left\n", " 565 entropy_left -= count_k * log(count_k)\n", " 566\n", " 567 count_k = sum_right[c]\n", " 568 if count_k > 0.0:\n", " 569 count_k /= self.weighted_n_right\n", " 570 entropy_right -= count_k * log(count_k)\n", " 571 #+-- 3 lines: sum_left += self.sum_stride----------------------------------------\n", " 574\n", " 575 impurity_left[0] = entropy_left / self.n_outputs\n", " 576 impurity_right[0] = entropy_right / self.n_outputs\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "#### 2.2 Gini\n", "$ \\text{ index } = 1 - \\sum_{k=0}^{K-1} \\text{ count_k}^2 $\n", "\n", "```Python\n", " 579 cdef class Gini(ClassificationCriterion):\n", " 580 \"\"\"Gini Index impurity criterion.\n", " 581 +-- 11 lines: This handles cases where the target is a classification taking valu\n", " 592 index = \\sum_{k=0}^{K-1} count_k (1 - count_k)\n", " 593 = 1 - \\sum_{k=0}^{K-1} count_k ** 2\n", " 594 \"\"\"\n", " 595\n", " 596 cdef double node_impurity(self) nogil:\n", " 597 #+-- 12 lines: \"\"\"Evaluate the impurity of the current node, i.e. the impurity of-\n", " 609 for k in range(self.n_outputs):\n", " 610 sq_count = 0.0\n", " 611\n", " 612 for c in range(n_classes[k]):\n", " 613 count_k = sum_total[c]\n", " 614 sq_count += count_k * count_k\n", " 615\n", " 616 gini += 1.0 - sq_count / (self.weighted_n_node_samples *\n", " 617 self.weighted_n_node_samples)\n", " 618\n", " 619 sum_total += self.sum_stride\n", " 620\n", " 621 return gini / self.n_outputs\n", " 622\n", " 623 cdef void children_impurity(self, double* impurity_left,\n", " 624 double* impurity_right) nogil:\n", " 625 \"\"\"Evaluate the impurity in children nodes\n", " 626 #+-- 45 lines: i.e. the impurity of the left child (samples[start:pos]) and the---\n", "```" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 3. 回归评价函数" ] }, { "cell_type": "markdown", "metadata": { "collapsed": true }, "source": [ "#### 3.0 RegressionCriterion\n", "RegressionCriterion 是回归评价的父类,与分类预计算权重占比 $p_k^i$ 类似,回归会预计算节点的方差值。为了节省计算量,对公式做了变形,降低了计算次数,具体推导如下所示:\n", "\n", "\\begin{align}\n", " \\text{var } &= \\frac{1}{n} \\sum_i^n (y_i - \\bar{y})^2 \\\\\n", " &= E \\left [ (y_i - E(y_i))^2 \\right ] \\\\\n", " &= E[y_i^2] - E^2[y_i] \\\\\n", " &= \\frac{1}{n} \\sum_i^n (y_i^2) - \\bar{y}^2 \\\\\n", " &= \\frac{1}{n} \\sum_i^n (y_i^2) - \\left ( \\frac{1}{n} \\sum_i^n y_i \\right )^2 \\\\\n", " & \\text{ 引入权重值 } \\\\\n", " &= \\sum_i^n ( w_i (y_i^2) ) / \\sum_i^n w_i - \\left ( \\sum_i^n (w_i y_i) / \\sum_i^n w_i \\right )^2 \\\\\n", " &= \\frac{\\text{sq_sum_total}}{\\text{weighted_n_node_samples}} - \\left (\\frac{\\text{sum_total}}{\\text{weighted_n_node_samples}} \\right )^2\n", "\\end{align}\n", "\n", "所以,RegressionCriterion 的方法都是在计算和更新上述式子中的三个变量(还有子节点的),因为整个代码和分类相似,不再赘述。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 3.1 MSE\n", "其实这里 MSE 用的公式就是上述的方差公式。虽然实际实义中 $\\text{MSE } = E [ y_i - \\hat{y}_i ]$,但因为决策树预测值用的是均值,所以全部 $\\hat{y}_i = \\bar{y_i}$,于是退化成了计算方差的公式。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "##### 3.2 FriedmanMSE\n", "FriedmanMSE 来源于 Friedman 论文 [Greedy Function Approximation: A Gradient Boosting Machine(https://statweb.stanford.edu/~jhf/ftp/trebst.pdf) 中的公式 (35),跟 Gradient Boosting 相关,这方面我还没看,等后继写到 GBDT 时再回来补上。" ] }, { "cell_type": "markdown", "metadata": {}, "source": [ "### 总结\n", "\n", "本文较为详细地介绍了 sklearn 中评估函数模块的实现细节:一个接口类,两个辅助计算的父类,四个具体计算的子类。" ] } ], "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": "3.5.2" } }, "nbformat": 4, "nbformat_minor": 0 }