머신러닝

[calmnis]

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

df = pd.read_csv('cust_data.csv')

df.info()

df.tail()

# termination 레이블 불균형 
df['termination'].value_counts().plot(kind='bar')

# Object 컬럼 리스트 정의
cal_cols = ['class', 'sex', 'stop', 'npay', 'termination', 'bill_rating']

# pandas get_dummies 함수 사용하여 Object 컬럼에 대해 One-Hot-Encoding 수행
df1 = pd.get_dummies(data=df, columns=cal_cols, drop_first=True)

df1.info()

X = df1.drop('termination_Y', axis=1).values

y = df1['termination_Y'].values

X.shape, y.shape

from sklearn.model_selection import train_test_split

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, stratify=y,random_state=42)

X_train.shape, y_train.shape

from sklearn.preprocessing import MinMaxScaler

scaler = MinMaxScaler()

X_train = scaler.fit_transform(X_train)
X_test = scaler.transform(X_test)

X_train[:2], y_train[:2]

from sklearn.metrics import accuracy_score

1. 로지스틱 회귀 (LogisticRegression, 분류)

from sklearn.linear_model import LogisticRegression
from sklearn.metrics import confusion_matrix 
from sklearn.metrics import accuracy_score, precision_score, recall_score, f1_score
from sklearn.metrics import classification_report

lg = LogisticRegression()
lg.fit(X_train, y_train)

lg.score(X_test, y_test)

lg_pred = lg.predict(X_test)

confusion_matrix(y_test, lg_pred) 

accuracy_score(y_test, lg_pred)  

precision_score(y_test, lg_pred) 

recall_score(y_test, lg_pred)  

f1_score(y_test, lg_pred) 

classification_report(y_test, lg_pred)

2. KNN (K-Nearest Neighbor)

from sklearn.neighbors import KNeighborsClassifier

knn = KNeighborsClassifier(n_neighbors=5)
knn.fit(X_train, y_train)

knn.score(X_test, y_test)

knn_pred = knn.predict(X_test)

confusion_matrix(y_test, knn_pred) 

accuracy_score(y_test, knn_pred) 

precision_score(y_test, knn_pred) 

recall_score(y_test, knn_pred) 

f1_score(y_test, knn_pred) 

classification_report(y_test, knn_pred) 

3. 결정트리(DecisionTree)

from sklearn.tree import DecisionTreeClassifier

dt = DecisionTreeClassifier(max_depth=10, random_state=42)
dt.fit(X_train, y_train)

dt.score(X_train, y_train)

dt_pred = dt.predict(X_test)

confusion_matrix(y_test, dt_pred) 

accuracy_score(y_test, dt_pred) 

precision_score(y_test, dt_pred) 

recall_score(y_test, dt_pred) 

f1_score(y_test, dt_pred) 

classification_report(y_test, dt_pred) 

4. 랜덤포레스트(RandomForest)

from sklearn.ensemble import RandomForestClassifier

rfc = RandomForestClassifier(n_estimators=3, random_state=42)
rfc.fit(X_train, y_train)

rfc.score(X_train, y_train)

rfc_pred = rfc.predict(X_test)

confusion_matrix(y_test, rfc_pred) 

accuracy_score(y_test, rfc_pred) 

precision_score(y_test, rfc_pred) 

recall_score(y_test, rfc_pred) 

f1_score(y_test, rfc_pred) 

classification_report(y_test, rfc_pred) 

5. XGBoost

!pip install xgboost

from xgboost import XGBClassifier

xgb = XGBClassifier(n_estimators=3, random_state=42)  
xgb.fit(X_train, y_train)

xgb.score(X_train, y_train)

xgb_pred = xgb.predict(X_test)

confusion_matrix(y_test, xgb_pred) 

accuracy_score(y_test, xgb_pred) 

precision_score(y_test, xgb_pred) 

recall_score(y_test, xgb_pred) 

f1_score(y_test, xgb_pred) 

classification_report(y_test, xgb_pred)