Flask_ML_Web서비스(3)_Cancer

conda create -n flask_house python=3.8

conda activate  flask_house

pip install -r requirements.txt

pip install flask==2.1.3  Jinja2

pip install scikit-learn  seaborn  pillow  imageio  pandas  virtualenv

pip install numpy==1.23.1

conda list

----[폴더명]

    ----- [폴더] static

    ----- [폴더] templates

    ----- cancer.py

    ----- flask_app.py

import pandas as pd

import pickle

import time

import sklearn as sk

from sklearn.model_selection import train_test_split

from sklearn.linear_model import LogisticRegression

from sklearn.ensemble import VotingClassifier

import lightgbm as lgbm

from lightgbm import LGBMClassifier

from sklearn.datasets import load_breast_cancer

start_time = time.time()

print("LGBM Version : ", lgbm.__version__)

print("Pandas Version : ", pd.__version__)

print("Scikit-Learn Version : ", sk.__version__)

cancer = load_breast_cancer()

cancer_df = pd.DataFrame(cancer.data, columns=cancer.feature_names)

print(cancer_df.columns)

# 피처와 레이블를 지정.

sel = ['worst texture', 'worst concave points', 'mean texture', 'area error', 'worst perimeter']

X = cancer_df[sel]

y = cancer.target

print("worst texture : (min max) : ", cancer_df['worst texture'].min(), cancer_df['worst texture'].max())

print("worst concave points : (min max) : ", cancer_df['worst concave points'].min(), cancer_df['worst concave points'].max())

print("mean texture : (min max) : ", cancer_df['mean texture'].min(), cancer_df['mean texture'].max())

print("area error : (min max) : ", cancer_df['area error'].min(), cancer_df['area error'].max())

print("worst perimeter : (min max) : ", cancer_df['worst perimeter'].min(), cancer_df['worst perimeter'].max())

X_train , X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

# 모델 선택

model_lgbm = LGBMClassifier(n_estimators= 400)

evals = [(X_test, y_test)]

model_lgbm.fit(X_train, y_train, early_stopping_rounds=100, eval_metric='logloss',

              eval_set=evals, verbose=True)

model_logit = LogisticRegression()

model_logit.fit(X_train, y_train)

print("lgbm 모델 정확도 : ", model_lgbm.score(X_test, y_test))

print("로지스틱회귀 모델 정확도 : ", model_logit.score(X_test, y_test))

vo_clf = VotingClassifier( estimators=[("logitlr", model_logit) ,

                                      ("lgbm", model_lgbm)] , voting='soft')

vo_clf.fit(X_train, y_train)

print("lgbm and 로지스틱회귀 앙상블 모델 정확도 : ", vo_clf.score(X_test, y_test))

pickle.dump(model_lgbm, open('cancer_lgbm_logit.pkl', 'wb'))

print("시간 (초): ", time.time() - start_time )