如何在Pytorch中进行火车测试分开

Question

使用Pytorch编码字符串值的最佳方法是什么？

df_train.head():
  country            league        home_team   away_team  home_odds  draw_odds  away_odds  home_score  away_score  dow  month
0  Brazil  Copa do Nordeste     Sport Recife  Imperatriz       1.36       4.31       7.66           2           2    4      2
1  Brazil  Copa do Nordeste              ABC  America RN       2.62       3.30       2.48           2           1    6      2
2  Brazil  Copa do Nordeste  Frei Paulistano     Nautico       5.19       3.58       1.62           0           2    6      2
3  Brazil  Copa do Nordeste      Botafogo PB   Confianca       2.06       3.16       3.50           1           1    6      2
4  Brazil  Copa do Nordeste        Fortaleza       Ceara       2.19       2.98       3.38           1           1    6      2

df_test.shape:
(76544, 11)

df_test.head()
     country          league      home_team      away_team  home_odds  draw_odds  away_odds  home_score  away_score  dow  month
0      World   Club Friendly       Westerlo           Gent       2.93       3.47       2.19         NaN         NaN    4      6
1   Malaysia    Super League       Johor DT       Selangor       1.27       5.59       8.26         NaN         NaN    4      6
2  Argentina  Reserve League        Lanus 2  River Plate 2       2.54       3.12       2.65         NaN         NaN    4      6
3       Asia         AFC Cup    Bali United          Kedah       1.58       4.08       4.93         NaN         NaN    4      6
4   Ethiopia  Premier League  Defence Force     Adama City       2.93       2.16       3.38         NaN         NaN    4      6

df_test.shape:
(599, 11)

我使用Pandas在Sklearn中进行编码：

def encode_features(df_train, df_test):
    features = ['country', 'league', 'home_team', 'away_team']
    df_combined = pd.concat([df_train[features], df_test[features]])

    for feature in features:
        le = preprocessing.LabelEncoder()
        le = le.fit(df_combined[feature])
        df_train[feature] = le.transform(df_train[feature])
        df_test[feature] = le.transform(df_test[feature])
    return df_train, df_test


df_train, df_test = encode_features(df_train, df_test)

Answer 1

为了用字符串编码这四列，您可以使用标签编码器或一个热编码器。这是使用标签编码器的情况的参考类。

import pandas
from sklearn.preprocessing import LabelEncoder

class MultiColumnLabelEncoder:
    def __init__(self,columns = None):
        self.columns = columns # array of column names to encode

    def fit(self,X,y=None):
        return self # not relevant here

    def transform(self,X):
        output = X.copy()
        if self.columns is not None:
            for col in self.columns:
                output[col] = LabelEncoder().fit_transform(output[col])
        else:
            for colname,col in output.iteritems():
                output[colname] = LabelEncoder().fit_transform(col)
        return output

    def fit_transform(self,X,y=None):
        return self.fit(X,y).transform(X)

MultiColumnLabelEncoder(columns = ['country', 'league', 'home_team', 'away_team']).fit_transform(df_train)

我认为这可能对您的情况有所帮助。