gcForest 官方代码详解

版权声明：本文为博主原创文章，未经博主允许不得转载。 https://blog.csdn.net/zwqjoy/article/details/84774888

1.介绍

gcForest v1.1.1是gcForest的一个官方托管在GitHub上的版本，是由Ji Feng(Deep Forest的paper的作者之一)维护和开发，该版本支持Python3.5,且有类似于Scikit-Learn的API接口风格，在该项目中提供了一些调用例子，目前支持的基分类器有RandomForestClassifier,XGBClassifer,ExtraTreesClassifier,LogisticRegression,SGDClassifier如果采用XGBoost的基分类器还可以使用GPU

本文采用的是v1.1.1版本，github地址https://github.com/kingfengji/gcForest

如果想增加其他基分类器，可以在模块中的lib/gcforest/estimators/__init__.py中添加

使用该模块需要依赖安装如下模块：

• argparse
• joblib
• keras
• psutil
• scikit-learn>=0.18.1
• scipy
• simplejson
• tensorflow
• xgboost

2.API调用样例

这里先列出gcForest提供的API接口：

• fit_tranform(X_train,y_train) 是gcForest模型最后一层每个估计器预测的概率concatenated的结果

• fit_transform(X_train,y_train,X_test=x_test,y_test=y_test) 测试数据的准确率在训练的过程中也会被记录下来

• set_keep_model_mem(False) 如果你的缓存不够，把该参数设置成False(默认为True),如果设置成False,你需要使用fit_transform(X_train,y_train,X_test=x_test,y_test=y_test)来评估你的模型

• predict(X_test) # 模型预测

• transform(X_test)

  扫描二维码关注公众号，回复： 4548047 查看本文章

代码主要分为两部分：examples文件夹下是主代码.py和配置文件.json；libs文件夹下是代码中用到的库

主代码的实现

最简单的调用gcForest的方式如下：

# 导入必要的模块
from gcforest.gcforest import GCForest

# 初始化一个gcForest对象
gc = GCForest(config) # config是一个字典结构

# gcForest模型最后一层每个估计器预测的概率concatenated的结果
X_train_enc = gc.fit_transform(X_train,y_train)

# 测试集的预测
y_pred = gc.predict(X_test)

lib库的详解

gcforest.py 整个框架的实现
fgnet.py 多粒度部分，FineGrained的实现
cascade/cascade_classifier 级联分类器的实现
datasets/.... 包含一系列数据集的定义
estimator/... 包含决策树在进行评估用到的函数（多种分类器的预估）
layer/... 包含不同的层操作，如连接、池化、滑窗等
utils/.. 包含各种功能函数，譬如计算准确率、win_vote、win_avg、get_windows等

json配置文件的详解

参数介绍

• max_depth: 决策树最大深度。默认为"None"，决策树在建立子树的时候不会限制子树的深度这样建树时，会使每一个叶节点只有一个类别，或是达到min_samples_split。一般来说，数据少或者特征少的时候可以不管这个值。如果模型样本量多，特征也多的情况下，推荐限制这个最大深度，具体的取值取决于数据的分布。常用的可以取值10-100之间。
• estimators表示选择的分类器
• n_estimators 为森林里的树的数量
• n_jobs: int (default=1)
  The number of jobs to run in parallel for any Random Forest fit and predict.
  If -1, then the number of jobs is set to the number of cores.

---

训练的配置，分三类情况：

1. 采用默认的模型

def get_toy_config():
    config = {}
    ca_config = {}
    ca_config["random_state"] = 0  # 0 or 1
    ca_config["max_layers"] = 100  #最大的层数，layer对应论文中的level
    ca_config["early_stopping_rounds"] = 3  #如果出现某层的三层以内的准确率都没有提升，层中止
    ca_config["n_classes"] = 3      #判别的类别数量
    ca_config["estimators"] = []  
    ca_config["estimators"].append(
            {"n_folds": 5, "type": "XGBClassifier", "n_estimators": 10, "max_depth": 5,
             "objective": "multi:softprob", "silent": True, "nthread": -1, "learning_rate": 0.1} )
    ca_config["estimators"].append({"n_folds": 5, "type": "RandomForestClassifier", "n_estimators": 10, "max_depth": None, "n_jobs": -1})
    ca_config["estimators"].append({"n_folds": 5, "type": "ExtraTreesClassifier", "n_estimators": 10, "max_depth": None, "n_jobs": -1})
    ca_config["estimators"].append({"n_folds": 5, "type": "LogisticRegression"})
    config["cascade"] = ca_config    #共使用了四个基学习器
    return config

支持的基本分类器：
RandomForestClassifier
XGBClassifier
ExtraTreesClassifier
LogisticRegression
SGDClassifier

你可以通过下述方式手动添加任何分类器：

lib/gcforest/estimators/__init__.py

1. 只有级联（cascade）部分

{
"cascade": {
    "random_state": 0,
    "max_layers": 100,
    "early_stopping_rounds": 3,
    "n_classes": 10,
    "estimators": [
        {"n_folds":5,"type":"XGBClassifier","n_estimators":10,"max_depth":5,"objective":"multi:softprob", "silent":true, "nthread":-1, "learning_rate":0.1},
        {"n_folds":5,"type":"RandomForestClassifier","n_estimators":10,"max_depth":null,"n_jobs":-1},
        {"n_folds":5,"type":"ExtraTreesClassifier","n_estimators":10,"max_depth":null,"n_jobs":-1},
        {"n_folds":5,"type":"LogisticRegression"}
    ]
}
}

1. “multi fine-grained + cascade” 两部分
   滑动窗口的大小： {[d/16], [d/8], [d/4]}，d代表输入特征的数量；
   "look_indexs_cycle": [
   [0, 1],
   [2, 3],
   [4, 5]]
   代表级联多粒度的方式，第一层级联0、1森林的输出，第二层级联2、3森林的输出，第三层级联4、5森林的输出

{
"net":{
"outputs": ["pool1/7x7/ets", "pool1/7x7/rf", "pool1/10x10/ets", "pool1/10x10/rf", "pool1/13x13/ets", "pool1/13x13/rf"],
"layers":[
// win1/7x7
    {
        "type":"FGWinLayer",
        "name":"win1/7x7",
        "bottoms": ["X","y"],
        "tops":["win1/7x7/ets", "win1/7x7/rf"],
        "n_classes": 10,
        "estimators": [
            {"n_folds":3,"type":"ExtraTreesClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10},
            {"n_folds":3,"type":"RandomForestClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10}
        ],
        "stride_x": 2,
        "stride_y": 2,
        "win_x":7,
        "win_y":7
    },
// win1/10x10
    {
        "type":"FGWinLayer",
        "name":"win1/10x10",
        "bottoms": ["X","y"],
        "tops":["win1/10x10/ets", "win1/10x10/rf"],
        "n_classes": 10,
        "estimators": [
            {"n_folds":3,"type":"ExtraTreesClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10},
            {"n_folds":3,"type":"RandomForestClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10}
        ],
        "stride_x": 2,
        "stride_y": 2,
        "win_x":10,
        "win_y":10
    },
// win1/13x13
    {
        "type":"FGWinLayer",
        "name":"win1/13x13",
        "bottoms": ["X","y"],
        "tops":["win1/13x13/ets", "win1/13x13/rf"],
        "n_classes": 10,
        "estimators": [
            {"n_folds":3,"type":"ExtraTreesClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10},
            {"n_folds":3,"type":"RandomForestClassifier","n_estimators":20,"max_depth":10,"n_jobs":-1,"min_samples_leaf":10}
        ],
        "stride_x": 2,
        "stride_y": 2,
        "win_x":13,
        "win_y":13
    },
// pool1
    {
        "type":"FGPoolLayer",
        "name":"pool1",
        "bottoms": ["win1/7x7/ets", "win1/7x7/rf", "win1/10x10/ets", "win1/10x10/rf", "win1/13x13/ets", "win1/13x13/rf"],
        "tops": ["pool1/7x7/ets", "pool1/7x7/rf", "pool1/10x10/ets", "pool1/10x10/rf", "pool1/13x13/ets", "pool1/13x13/rf"],
        "pool_method": "avg",
        "win_x":2,
        "win_y":2
    }
]

},

"cascade": {
    "random_state": 0,
    "max_layers": 100,
    "early_stopping_rounds": 3,
    "look_indexs_cycle": [
        [0, 1],
        [2, 3],
        [4, 5]
    ],
    "n_classes": 10,
    "estimators": [
        {"n_folds":5,"type":"ExtraTreesClassifier","n_estimators":1000,"max_depth":null,"n_jobs":-1},
        {"n_folds":5,"type":"RandomForestClassifier","n_estimators":1000,"max_depth":null,"n_jobs":-1}
    ]
}
}

3.MNIST样例

下面我们使用MNIST数据集来演示gcForest的使用及代码的详细说明：

# 导入必要的模块

import argparse # 命令行参数调用模块
import numpy as np 
import sys
from keras.datasets import mnist # MNIST数据集
import pickle
from sklearn.ensemble import RandomForestClassifier
from sklearn.metrics import accuracy_score
sys.path.insert(0, "lib")

from gcforest.gcforest import GCForest
from gcforest.utils.config_utils import load_json


def parse_args():
	'''
	解析终端命令行参数(model)
	'''
    parser = argparse.ArgumentParser()
    parser.add_argument("--model", dest="model", type=str, default=None, 
	help="gcfoest Net Model File")
    args = parser.parse_args()
    return args


def get_toy_config():
	'''
	生成级联结构的相关结构
	'''
    config = {}
    ca_config = {}
    ca_config["random_state"] = 0
    ca_config["max_layers"] = 100
    ca_config["early_stopping_rounds"] = 3
    ca_config["n_classes"] = 10
    ca_config["estimators"] = []
    ca_config["estimators"].append(
            {"n_folds": 5, "type": "XGBClassifier", "n_estimators": 10, 
		"max_depth": 5,"objective": "multi:softprob", "silent": 
		True, "nthread": -1, "learning_rate": 0.1} )
    ca_config["estimators"].append({"n_folds": 5, "type": "RandomForestClassifier", 
	"n_estimators": 10, "max_depth": None, "n_jobs": -1})
    ca_config["estimators"].append({"n_folds": 5, "type": "ExtraTreesClassifier",
	 "n_estimators": 10, "max_depth": None, "n_jobs": -1})
    ca_config["estimators"].append({"n_folds": 5, "type": "LogisticRegression"})
    config["cascade"] = ca_config
    return config

# get_toy_config()生成的结构，如下所示：

'''
{
"cascade": {
    "random_state": 0,
    "max_layers": 100,
    "early_stopping_rounds": 3,
    "n_classes": 10,
    "estimators": [
        {"n_folds":5,"type":"XGBClassifier","n_estimators":10,"max_depth":5,
		"objective":"multi:softprob", "silent":true, 
		"nthread":-1, "learning_rate":0.1},
        {"n_folds":5,"type":"RandomForestClassifier","n_estimators":10,
		"max_depth":null,"n_jobs":-1},
        {"n_folds":5,"type":"ExtraTreesClassifier","n_estimators":10,
		"max_depth":null,"n_jobs":-1},
        {"n_folds":5,"type":"LogisticRegression"}
    ]
}
}
'''

if __name__ == "__main__":
    args = parse_args()
    if args.model is None:
        config = get_toy_config()
    else:
        config = load_json(args.model)

    gc = GCForest(config)
    # 如果模型消耗太大内存，可以使用如下命令使得gcforest不保存在内存中
    # gc.set_keep_model_in_mem(False), 默认情况下是True.

    (X_train, y_train), (X_test, y_test) = mnist.load_data()
    # X_train, y_train = X_train[:2000], y_train[:2000]
    # np.newaxis相当于增加了一个维度
    X_train = X_train[:, np.newaxis, :, :]
    X_test = X_test[:, np.newaxis, :, :]


    X_train_enc = gc.fit_transform(X_train, y_train)
    # X_enc是gcForest模型最后一层每个估计器预测的概率concatenated的结果
    # X_enc.shape =
    #   (n_datas, n_estimators * n_classes): 如果是级联结构
    #   (n_datas, n_estimators * n_classes, dimX, dimY): 如果只有多粒度扫描结构

    # 可以在fit_transform方法中加入X_test,y_test,这样测试数据的准确率在训练的过程中
    # 也会被记录下来。
    # X_train_enc, X_test_enc = 
	gc.fit_transform(X_train, y_train, X_test=X_test, y_test=y_test)

    # 注意: 如果设置了gc.set_keep_model_in_mem(True),必须使用
    # gc.fit_transform(X_train, y_train, X_test=X_test, y_test=y_test)
    # 评估模型

    # 测试集预测与评估
    y_pred = gc.predict(X_test)
    acc = accuracy_score(y_test, y_pred)
    print("Test Accuracy of GcForest = {:.2f} %".format(acc * 100))

    # 可以使用gcForest得到的X_enc数据进行其他模型的训练比如xgboost/RF
    # 数据的concat
    X_test_enc = gc.transform(X_test)
    X_train_enc = X_train_enc.reshape((X_train_enc.shape[0], -1))
    X_test_enc = X_test_enc.reshape((X_test_enc.shape[0], -1))
    X_train_origin = X_train.reshape((X_train.shape[0], -1))
    X_test_origin = X_test.reshape((X_test.shape[0], -1))
    X_train_enc = np.hstack((X_train_origin, X_train_enc))
    X_test_enc = np.hstack((X_test_origin, X_test_enc))

    print("X_train_enc.shape={}, X_test_enc.shape={}".format(X_train_enc.shape,
	 X_test_enc.shape))

    # 训练一个RF
    clf = RandomForestClassifier(n_estimators=1000, max_depth=None, n_jobs=-1)
    clf.fit(X_train_enc, y_train)
    y_pred = clf.predict(X_test_enc)
    acc = accuracy_score(y_test, y_pred)
    print("Test Accuracy of Other classifier using 
	gcforest's X_encode = {:.2f} %".format(acc * 100))

    # 模型写入pickle文件
    with open("test.pkl", "wb") as f:
        pickle.dump(gc, f, pickle.HIGHEST_PROTOCOL)

    # 加载训练的模型
    with open("test.pkl", "rb") as f:
        gc = pickle.load(f)
    y_pred = gc.predict(X_test)
    acc = accuracy_score(y_test, y_pred)
    print("Test Accuracy of GcForest (save and load) = {:.2f} %".format(acc * 100))

这里需要注意的是gcForest不但可以对传统的结构化的2维数据建模，还可以对非结构化的数据比如图像，序列化的文本数据，音频数据等进行建模，但要注意数据维度的设定：

• 如果仅使用级联结构，X_train,X_test对于2-D数组其维度为(n_samples,n_features);3-D或4-D数组会自动reshape为2-D，例如MNIST数据(60000,28,28)会reshape为(60000,784),(60000,3,28,28)会reshape为(60000,2352)。

• 如果使用多粒度扫描结构，X_train,X_test必须是4—D的数组,图像数据其维度是(n_samples,n_channels,n_height,n_width)；序列数据其维度为(n_smaples,n_features,seq_len,1),例如对于IMDB数据，n_features为1，对于音频MFCC特征，其n_features可以为13,26等。

上述代码可以通过两种方式运行：

• 一种方式是通过json文件定义模型结构，比如级联森林结构，只需要写一个json文件如代码中显示的结构，然后通过命令行运行python examples/demo_mnist.py --model examples/demo_mnist-gc.json就可以完成训练；如果既使用多粒度扫面又使用级联结构，那么需要同时把多粒度扫描的结构定义出来。
• 定义好的json可以通过模块中的load_json()方法加载，然后作为参数初始化模型，如下：

config = load_json(your_json_file)
gc = GCForest(config) 

• 另一种方式是直接通过Python代码定义模型结构，实际上模型结构就是一个字典数据结构，即是上述代码中的get_toy_config()方法。