取相關係數大於0.3的決策樹baseline
阿新 • • 發佈：2022-12-09
模型在測試集的準確率為0.74提升了一些說明根據相關係數取模型是不錯的選擇。
import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import seaborn as sns
df = pd.read_csv('train.csv')
df=df.drop(['ID'],axis=1)
df=df.to_numpy()
feature=np.abs(np.fft.fft(df[:,:-1]))
feature=np.concatenate((feature,np.reshape(df[:,-1],(-1,1))),axis=1)
train=pd.DataFrame(feature)
heat=train.corr()
fe=heat.index[abs(heat[240])>0.3]
train=train.to_numpy()
train=train[:,fe]
from sklearn.model_selection import train_test_split
from sklearn.metrics import accuracy_score
from sklearn import tree
from sklearn.model_selection import cross_val_score
from sklearn.model_selection import KFold
kf=KFold(n_splits=5,shuffle=False)
for k in range(30):
sum=0
sum1=0
i=0
for train_index,test_index in kf.split(train):
i=i+1
tfeature=train[train_index,:-1]
label=train[train_index,-1]
clf=tree.DecisionTreeClassifier(criterion='entropy',random_state=0,max_depth=k+1)
clf.fit(tfeature,label)
l=clf.predict(tfeature)
ttest=train[test_index,:-1]
testlabel=train[test_index,-1]
l1=clf.predict(ttest)
pr=accuracy_score(label, l)
pr1=accuracy_score(testlabel, l1)
sum=sum+pr
sum1=sum1+pr1
clf1=tree.DecisionTreeClassifier(criterion='entropy',random_state=0,max_depth=k+1)
scores = cross_val_score(clf1, train[:,:-1], train[:,-1], cv=5)
print(k,sum/i,sum1/i,scores.mean())
clf1=tree.DecisionTreeClassifier(criterion='entropy',random_state=0,max_depth=4+1)
clf1.fit(train[:,:-1],train[:,-1])
df1 = pd.read_csv('test.csv')
df1=df1.drop(['ID'],axis=1)
df1=df1.to_numpy()
feature=np.abs(np.fft.fft(df1[:,:]))
feature=feature[:,fe[:-1]]
out=clf1.predict(feature)
out=pd.DataFrame(out)
out.columns = ['CLASS']
w=[]
for k in range(out.shape[0]):
w.append(k+210)
out['ID']=np.reshape(w,(-1,1))
out[['ID','CLASS']].to_csv('out3.csv',index=False)