deepctr deepfm 多分类实现还没结果，但是也没有bug

Discussion

Home Discussion Board High-Energy Particle Classification Challenge deepctr deepfm 多分类实现还没结果，但是也没有bug

空空

deepctr deepfm 多分类实现还没结果，但是也没有bug

posted in High-Energy Particle Classification Challenge

2019-12-22 09:00

```python
import pandas as pd
from sklearn.metrics import log_loss, roc_auc_score
from sklearn.model_selection import train_test_split
from sklearn.preprocessing import LabelEncoder, MinMaxScaler

from deepctr.models import DeepFM
from deepctr.inputs import SparseFeat, DenseFeat, get_feature_names

if __name__ == "__main__":
    data = pd.read_csv('jet_simple_data/simple_train_R04_jet.csv')

# jet_id,number_of_particles_in_this_jet,jet_px,jet_py,jet_pz,jet_energy,jet_mass,event_id,label
    # ts 单值类别特征
    sparse_features = ["event_id"]
    # 数值特征
    dense_features = ["number_of_particles_in_this_jet", "jet_px", "jet_py", "jet_pz", "jet_energy", "jet_mass"]

data[sparse_features] = data[sparse_features].fillna('-1', )
    data[dense_features] = data[dense_features].fillna(0, )
    target = ['label']
    #
    # 1.稀疏特征的标签编码，密集特征的简单变换
    for feat in sparse_features:
        lbe = LabelEncoder()
        data[feat] = lbe.fit_transform(data[feat])
    mms = MinMaxScaler(feature_range=(0, 1))
    data[dense_features] = mms.fit_transform(data[dense_features])

# 2.统计每个稀疏字段的唯一特征，并记录密集特征字段名

fix_len_feature_columns = [SparseFeat(feat, vocabulary_size=data[feat].nunique(), embedding_dim=4)
                               for i, feat in enumerate(sparse_features)] + [DenseFeat(feat, 1, )
                                                                            for feat in dense_features]

dnn_feature_columns = fix_len_feature_columns
    linear_feature_columns = fix_len_feature_columns

feature_names = get_feature_names(linear_feature_columns + dnn_feature_columns)

# 3.生成模型的输入数据

train, test = train_test_split(data, test_size=0.2)
    train_model_input = {name: train[name] for name in feature_names}
    test_model_input = {name: test[name] for name in feature_names}

# 4.定义模型、训练、预测和评估
    model = DeepFM(linear_feature_columns, dnn_feature_columns, task='binary')
    model.compile("adam", "sparse_categorical_crossentropy",
                  metrics=['sparse_categorical_crossentropy','accuracy'], )

history = model.fit(train_model_input, train[target].values,
                        batch_size=4096, epochs=100, verbose=1, validation_split=0.2, )
    pred_ans = model.predict(test_model_input, batch_size=4096)
    print("test LogLoss", round(log_loss(test[target].values, pred_ans), 4))
    print("test AUC", round(roc_auc_score(test[target].values, pred_ans), 4))

```

23 comments