加载 pickle 模型以预测另一个数据集

Question

我有两个数据集。第一个是 train ，另一个是 test 。首先，我使用 LogisticRegression 来训练我的模型并保存为 ;

df = pd.read_csv('train.csv')

X = df[['Age', 'Gender']]
y = df['Race']

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


ohe = OneHotEncoder(handle_unknown='ignore')
ohe.fit(X_train)
X_train_ohe = ohe.transform(X_train).toarray()
ohe_df = pd.DataFrame(X_train_ohe, columns=ohe.get_feature_names(X_train.columns))
clf= LogisticRegression(solver='liblinear')
clf.fit(X_train_ohe, y_train.values.ravel())
X_test_ohe = ohe.transform(X_test)
y_preds = clf.predict(X_test_ohe)


print(" Logistic Regression ")
print ("Logistic Regression : ",accuracy_score(y_test, y_preds))
print(classification_report(y_test, y_preds)) 



model = LogisticRegression(solver='liblinear')
model.fit(X_train, y_train)


pickle.dump(model,open('model.pkl','wb'))

在这一步，我想创建一个新的 python 文件并读取 test.csv 文件。但是，test.csv 在“Race”标头中有 NaN 值。我可以在 test.csv 中用 0 填充 NaN 值，但是在这一步之后我必须做什么？我必须使用加载的模型来预测 Race 值。我试过这样的代码块：

df = pd.read_csv('test.csv')
df['Race']=df['Race'].replace(['NaN'], '0')
X = df[['Age', 'Gender']]


model=pickle.load(open('model.pkl','rb'))

Answer 1

model=pickle.load(open('model.pkl','rb'))
df = pd.read_csv('test.csv')
predictions=model.predict(df.drop('Race',axis=1))
print(predictions)