前言：Python在统计、绘图方面也有较多的库，有更强的扩展性，且机器学习目前主要为Python语言。源代码的意义在于调参。像很多科研绘图的大小、像素、色彩，还有机器学习的耦合效率，都需要经过调参，而这需要在源代码上调试。
import numpy as npimport pandas as pdimport matplotlib.pyplot as pltimport seaborn as snsplt.rc('font',family='Times New Roman')a = 'C:/Users/46685/Desktop/科研数据/数据汇总/新建处理后后连续变量.xlsx'b= pd.read_excel(a,sheet_name = 'Sheet1')b.head()
b.describe() #b为数据np.std(b)**2a= pd.read_excel(filePath_01,sheet_name = 'Sheet1')ls1 = a["CA724"]ls2 = a["年龄"]data = pd.DataFrame({'CA724':ls1,'年龄':ls2 })# 首先绘制出各属性关系图sns.pairplot(data,kind='scatter',diag_kind='kde')for column in data.columns:u = data[column].mean() # 计算均值std = data[column].std() # 计算标准差r,p = scipy.stats.kstest(data[column],'norm',(u,std))if p>0.05:print('拒绝原假设，显著性水平为{}，变量{}服从正态分布'.format(p,column))else:print('接受原假设，显著性水平为{}，变量{}不服从正态分布'.format(p,column))
from scipy import statsstats.pearsonr(b.胆总管扩张,b.肿块最大直径)
sns.scatterplot(x=b["胆总管扩张"], y=b["肿块硬度"])sns.jointplot(x='BMI',y='Waist',data=b,kind='reg',height=5,color='green')sns.jointplot(x='BMI',y='Waist',data=b,kind='hex',height=5)sns.jointplot(x='BMI',y='Waist',data=b,kind='kde',height=5)plt.figure(figsize=(10,8), dpi= 80)sns.pairplot(b, kind="scatter", plot_kws=dict(s=80, edgecolor="white", linewidth=2.5))plt.show()
df_coor=df.corr()df_coor.head()plt.subplots(figsize=(14,14),dpi=100,facecolor='w')# 设置画布大小，分辨率，和底色fig=sns.heatmap(df_coor,annot=True, vmax=1, square=True, cmap="Blues", fmt='.3f')#annot为热力图上显示数据；fmt='.2f'为数据保留小数点后两位,square呈现正方形，vmax最大值为1figfig.get_figure().savefig('df_corr.png',bbox_inches='tight',transparent=True)#保存图片#bbox_inches让图片显示完整，transparent=True让图片背景透明
from feature_selector import FeatureSelectorfs = FeatureSelector( data= x, labels = y)fs.identify_collinear(correlation_threshold=0.8, one_hot=False)correlated_features = fs.ops['collinear']fs.identify_zero_importance(task = 'classification',eval_metric = 'auc',n_iterations = 100, #n_iterations：模型训练的迭代次数；最终的特征重要性是n次迭代的平均值；early_stopping = False) # early_stopping: True/False, 是否需要提前停止# list of zero importance featureszero_importance_features = fs.ops['zero_importance']lw = 2plt.figure (figsize=(5,5),dpi=1000)plt.rc('font',family='Times New Roman')fs.plot_feature_importances(threshold = 0.9, plot_n = 14 )plt.show()plt.close()
from sklearn.model_selection import train_test_splitx_train,x_test,y_train,y_test=train_test_split(x,y,test_size=0.15,random_state=0)from sklearn.tree import DecisionTreeClassifierdtree = DecisionTreeClassifier(criterion='entropy',max_depth=5)dtree.fit(x_train,y_train)y_predict = dtree.predict(x_test)from sklearn.tree import plot_treefrom sklearn.metrics import classification_reportprint(classification_report(y_test,y_predict)) # 输出相关结果的函数import matplotlib.pyplot as pltfig = plt.figure(figsize=(16,12))a = plot_tree(dtree, feature_names=x.columns, fontsize=12, filled=True,class_names=['0', '1'])
