【Kaggle】 Digit Recognizer 详解

Kaggle : https://www.kaggle.com/c/digit-recognizer/

准确率：0.966

分类器：随机森林

代码详解：

import pandas as pd


# Random Forest Classifier
def random_forest_classifier(train_x, train_y):
	from sklearn.ensemble import RandomForestClassifier
	#使用RF分类器模型，定义三个参数值n_estimators=400, n_jobs=4, verbose=1
	model = RandomForestClassifier(n_estimators=400, n_jobs=4, verbose=1)
	model.fit(train_x, train_y)
	return model

#数据读取
def read_data(data_file):
	data = pd.read_csv(data_file)
	train = data[:int(len(data) * 1)]
	test = pd.read_csv("/Users/Cheney/Downloads/kaggle(方老师)/test.csv")
	#将标签和features分开
	train_y = train.label
	train_x = train.drop('label', axis=1)
	return train_x, train_y, test


if __name__ == '__main__':
	data_file = "/Users/Cheney/Downloads/kaggle(方老师)/train.csv"
	#用模型进行训练
	train_x, train_y, test = read_data(data_file)
	model = random_forest_classifier(train_x, train_y)
	#预测数据
	predict = model.predict(test)
	#数据写出
	dataframe = pd.DataFrame(predict)
	dataframe.to_csv('/Users/Cheney/Downloads/kaggle(方老师)/res.csv')