Untitled

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
import sklearn

df = pd.read_csv("output.csv")

df.head(100)

sns.displot(df['ft'])
sns.displot(df['se'])

sns.jointplot(x=df['ft'], y=df['se'])

X=df.drop(['se'],axis=1)
y=df['se']

best = 0
score = 0

from sklearn.model_selection import train_test_split
from sklearn.linear_model import LogisticRegression
from sklearn.metrics import confusion_matrix,accuracy_score, recall_score, precision_score

for i in range(0,1000,10):
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=i)
    reg = LogisticRegression(solver='lbfgs', max_iter=1000)
    reg.fit(X_train, y_train)
    predictions = reg.predict(X_test)
    if accuracy_score(y_test, predictions) > score:
        best = i
        score = accuracy_score(y_test, predictions)
    #print(best, score)


X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=best)
reg = LogisticRegression(solver='lbfgs', max_iter=1000)
reg.fit(X_train, y_train)
predictions = reg.predict(X_test)
print(predictions)


cnt_r=0
cnt_p=0


for i in range(0, 30):
  if X_test['re'].iloc[i] == 0:
    cnt_r+=1
    
  if(predictions[i] == 2 and X_test['ft'].iloc[i] == 0):
    cnt_p+=1
  elif(predictions[i]+1 == X_test['ft'].iloc[i]):
    cnt_p+=1



cnt_r /= 30
cnt_p /= 30


print(f"最佳參數:{best} 正確率: {score}")
print(f"預測勝率:{cnt_p}, 實際勝率:{cnt_r}")

result_df = X_test.copy()
result_df['True'] = y_test.values
result_df['Predicted'] = predictions

print(result_df)