빅데이터 분석 기사 공부 기출문제 4회 고객 Seg 분류 문제

 
 

#고객이 속한 세그멘트를 예측하고, 
#해당 예측 결과를 csv 파일로 저장하시오

import pandas as pd


X_test = pd.read_csv('https://raw.githubusercontent.com/YoungjinBD/dataset/main/CS_Seg_X_test.csv')
X_train = pd.read_csv('https://raw.githubusercontent.com/YoungjinBD/dataset/main/CS_Seg_X_train.csv')
y_train = pd.read_csv('https://raw.githubusercontent.com/YoungjinBD/dataset/main/CS_Seg_y_train.csv')

#print(X_train.head())

#       ID  Gender Ever_Married  Age Graduated  Profession  Work_Experience Spending_Score  Family_Size
#0  462809    Male           No   22        No  Healthcare              1.0            Low          4.0
#1  466315  Female          Yes   67       Yes    Engineer              1.0            Low          1.0
#2  461735    Male          Yes   67       Yes      Lawyer              0.0           High          2.0
#3  461319    Male          Yes   56        No      Artist              0.0        Average          2.0
#4  460156    Male           No   32       Yes  Healthcare              1.0            Low          3.0

# null이 있는지 확인하자
# 항상 보면 train는 세트로 있고, test는 세트로 없다
print(f' X_train : {X_train.isnull().sum().sum()}')
print(f' y_train : {X_train.isnull().sum().sum()}')
print(f' X_test : {X_train.isnull().sum().sum()}')

# 데이터 나누기
# 범주형 데이터
X_train_word = X_train[['Gender','Ever_Married','Graduated','Profession','Spending_Score']]
X_test_word = X_test[['Gender','Ever_Married','Graduated','Profession','Spending_Score']]
#print(X_train_word )

#수치형 데이터
X_train_num = X_train.drop(columns=['ID','Gender','Ever_Married','Graduated',
                                    'Profession','Spending_Score'])
X_test_num = X_train.drop(columns=['ID','Gender','Ever_Married','Graduated',
                                    'Profession','Spending_Score'])
print(X_train_num )

#데이터 스케일링
from sklearn.preprocessing import MinMaxScaler

#MinMax 스케일러 생성
scaler = MinMaxScaler()

#선택한 특성에 MinMax 스케일러를 적용하고 데이터 전환
X_train_num_scale = scaler.fit_transform(X_train_num)
X_test_num_scale = scaler.transform(X_test_num)

#테이터 프레임 설정
df_train_num = pd.DataFrame(X_train_num_scale, columns = X_train_num.columns)
df_test_num = pd.DataFrame(X_test_num_scale, columns = X_test_num.columns)

# 원핫 인코딩
df_train_word = pd.get_dummies(X_train_word)
df_test_word = pd.get_dummies(X_test_word)

#데이터결합
df_train = pd.concat([df_train_num, df_train_word], axis= 1)
df_test = pd.concat([df_test_num, df_train_word], axis= 1)
print(df_train)

from pandas.core.common import random_state
import xgboost as xgb
from sklearn.model_selection import train_test_split
from sklearn.metrics import classification_report

#target 값 변경
# 특정 항목값 변경하기
y_train[y_train['Segmentation'] == 'A'] = 0
y_train[y_train['Segmentation'] == 'B'] = 1
y_train[y_train['Segmentation'] == 'C'] = 2
y_train[y_train['Segmentation'] == 'D'] = 3

#모델 생성
model = xgb.XGBClassifier(random_state=77)

#Train, validation 데이터 설정
X_train, X_val, y_train, y_val = train_test_split(df_train.values, y_train['Segmentation'].values
,test_size=0.3)

#모델 학습
model.fit(X_train, y_train_t)

#validation 데이터로 성능 평가
y_pred = model.predict(X_val)
print(classification_report(y_val.astype(int), y_pred))

#예측
y_pred = model.predict(df_test)
df = pd.DataFrame(y_train['ID'], columns=['ID'])
df['Segmentation'] = y_pred

# Segmentation 데이터를 숫자에서 문자로 수정
df['Segmentation'][df['Segmentation']==0] = 'A'
df['Segmentation'][df['Segmentation']==0] = 'B'
df['Segmentation'][df['Segmentation']==0] = 'C'
df['Segmentation'][df['Segmentation']==0] = 'D'

#CSV 파일 저장
df.to_csv('result.csv', index=False)