# 公众号 Python实用宝典
import autograd.numpy as np
from autograd import grad
def oneline(x):
y = x/2
return y
grad_oneline = grad(oneline)
print(grad_oneline(3.0))
运行代码，传入任意X值，你就能得到在该X值下的斜率：
(base) G:\push\20220724>python 1.py
0.5
由于这是一条直线，因此无论你传什么值，都只会得到0.5的结果。
那么让我们再试试一个tanh函数：
# 公众号 Python实用宝典
import autograd.numpy as np
from autograd import grad
def tanh(x):
y = np.exp(-2.0 * x)
return (1.0 - y) / (1.0 + y)
grad_tanh = grad(tanh)
print(grad_tanh(1.0))
def training_loss(weights):
# Training loss is the negative log-likelihood of the training labels.
preds = logistic_predictions(weights, inputs)
label_probabilities = preds * targets + (1 - preds) * (1 - targets)
return -np.sum(np.log(label_probabilities))
# Define a function that returns gradients of training loss using Autograd.
training_gradient_fun = grad(training_loss)
# Optimize weights using gradient descent.
weights = np.array([0.0, 0.0, 0.0])
print("Initial loss:", training_loss(weights))
for i in range(100):
weights -= training_gradient_fun(weights) * 0.01
print("Trained loss:", training_loss(weights))
from sklearn.preprocessing import LabelEncoder
var_mod = ['Airline','Source','Destination','Additional_Info','Total_Stops','weekday','month','Dep_Time']
le = LabelEncoder()
for i in var_mod:
flights[i] = le.fit_transform(flights[i])
flights.head()
对每列数据进行特征缩放，提取自变量（x）和因变量（y）：
flights.corr()
def outlier(df):
for i in df.describe().columns:
Q1=df.describe().at['25%',i]
Q3=df.describe().at['75%',i]
IQR= Q3-Q1
LE=Q1-1.5*IQR
UE=Q3+1.5*IQR
df[i]=df[i].mask(df[i]<LE,LE)
df[i]=df[i].mask(df[i]>UE,UE)
return df
flights = outlier(flights)
x = flights.drop('Price',axis=1)
y = flights['Price']
划分测试集和训练集：
from sklearn.model_selection import train_test_split
x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, random_state=101)
4.2 模型训练及测试
使用随机森林进行模型训练：
from sklearn.ensemble import RandomForestRegressor
rfr=RandomForestRegressor(n_estimators=100)
rfr.fit(x_train,y_train)
def jieba_cut_and_save_file(inputList, n_weight, a_weight, output_cleaned_file=False):
"""
1. 读取中文文件并分词句子
2. 可以将分词后的结果保存到文件
"""
output_file = os.path.join('./data/', 'cleaned_' + 'trainMatrix.txt')
lines = []
tags = []
for line in inputList:
result = pseg.cut(clean_str(line))
a = []
b = []
for word, flag in result:
# 对分词后的新闻
if word != ' ':
# 若非空
a.append(word)
if flag.find('n')==0:
# 若是名词
b.append(n_weight)
elif flag.find('a')==0:
# 若形容词
b.append(a_weight)
else:
b.append(1)
lines.append(a)
tags.append(b)
if output_cleaned_file:
with open(output_file, 'w') as f:
for line in lines:
f.write(" ".join(line) + '\n')
vocabulary = createVocabList(lines)
# 根据词典生成词向量化器,并进行词向量化
setOfWords2Vec = setOfWords2VecFactory(vocabulary)
vectorized = []
for i,news in enumerate(lines):
vector = setOfWords2Vec(news, tags[i])
vectorized.append(vector)
return vectorized, vocabulary
def jieba_cut_and_save_file(inputList, output_cleaned_file=False):
"""
1. 读取中文文件并分词句子
2. 可以将分词后的结果保存到文件
3. 如果已经存在经过分词的数据文件则直接加载
"""
output_file = os.path.join('./data/', 'cleaned_' + 'trainMatrix.txt')
if os.path.exists(output_file):
lines = list(open(output_file, 'r').readlines())
lines = [line.strip('\n').split(' ') for line in lines]
else:
lines = [list(jieba.cut(clean_str(line))) for line in inputList]
# 将句子进行clean_str处理后进行结巴分词
lines = [[word for word in line if word != ' '] for line in lines]
if output_cleaned_file:
with open(output_file, 'w') as f:
for line in lines:
f.write(" ".join(line) + '\n')
vocabulary = createVocabList(lines)
# 根据词典生成词向量化器,并进行词向量化
setOfWords2Vec = setOfWords2VecFactory(vocabulary)
vectorized = [setOfWords2Vec(news) for news in lines]
return vectorized, vocabulary
def clean_str(string):
"""
1. 将除汉字外的字符转为一个空格
2. 除去句子前后的空格字符
"""
string = re.sub(r'[^\u4e00-\u9fff]', ' ', string)
string = re.sub(r'\s{2,}', ' ', string)
return string.strip()
def setOfWords2VecFactory(vocabList):
"""
通过给定词典，构造该词典对应的setOfWords2Vec
"""
#优化：通过事先构造词语到索引的哈希表，加快转化
index_map = {}
for i, word in enumerate(vocabList):
index_map[word] = i
def setOfWords2Vec(news):
"""
以在构造时提供的词典为基准词典向量化一条新闻
"""
result = [0]*len(vocabList)
for word in news:
#通过默认值查询同时起到获得索引与判断有无的作用
index = index_map.get(word, None)
if index:
result[index] = 1
return result
return setOfWords2Vec
5.最后是测试需要用到的向量化新闻
原理和前边对训练集的向量化相似，但是由于是测试时对单个新闻的向量化，因此我们把它分开了。
def vectorize_newslist(news_list, vocabulary):
"""
将新闻列表新闻向量化，变成词向量矩阵
注：如果没有词典，默认值为从集合中创造
"""
# 分词与过滤
cut_news_list = [list(jieba.cut(clean_str(news))) for news in news_list]
# 根据词典生成词向量化器,并进行词向量化
setOfWords2Vec = setOfWords2VecFactory(vocabulary)
vectorized = [setOfWords2Vec(news) for news in cut_news_list]
return vectorized, vocabulary
