美购精排模型

mlp/train.py

@@ -188,5 +188,6 @@ if __name__ == '__main__':
 
 
     predict(model_file,df_test)
+
     pass
 

spark/featureEng.py

@@ -212,7 +212,7 @@ def getDataVocab(samples):
 
     for c in samples.columns:
         # 判断是否以Bucket结尾 和 类别特征
-        if c.endswith("Bucket") or c in ITEM_CATE_COLUMNS:
+        if c.endswith("Bucket") or c in ITEM_CATE_COLUMNS+["userid","itemid"]:
             datas = samples.select(c).distinct().collect()
             vocabSet = set()
             for d in datas: