neurobcl package

Subpackages

• neurobcl.base package
  • Submodules
  • neurobcl.base.model module
    • NeuroBucketClassifier
      • NeuroBucketClassifier.fromJson()
      • NeuroBucketClassifier.get()
      • NeuroBucketClassifier.toJson()
    • NeuroBucketTrainer
      • NeuroBucketTrainer.depth_features_index()
      • NeuroBucketTrainer.get_at()
      • NeuroBucketTrainer.get_bucket_features()
      • NeuroBucketTrainer.get_non_bucket_features()
      • NeuroBucketTrainer.index()
      • NeuroBucketTrainer.total_items()
    • order_invariant()
  • Module contents
• neurobcl.trainers package
  • Submodules
  • neurobcl.trainers.dictionary_trainer module
    • DictionaryBucketTrainer
      • DictionaryBucketTrainer.get_at()
      • DictionaryBucketTrainer.get_bucket_features()
      • DictionaryBucketTrainer.get_non_bucket_features()
      • DictionaryBucketTrainer.total_items()
  • Module contents

Submodules

neurobcl.main module

neurobcl.main.train_from_dictionary(data: List[Dict], keyword_feats_name, bucket_feats_name, quantile_gap=10, max_depth=2) → NeuroBucketClassifier

Train using dictionary/json data, the format of the data should be a list of dictionaries where each dictionary represents a single item and the keys of the dictionary represent the features of the item. (Should be uniform)

Parameters:

• data (List[Dict]) – The data to train on

• keyword_feats_name (List[str]) – The name of the keyword features (categorical features)

• bucket_feats_name (List[str]) – The name of the bucket features (numerical features)

• quantile_gap (int, optional) – The gap between the quantiles

• max_depth (int, optional) – The maximum depth of the decision tree

Returns:

The trained model

Return type:

neurobcl.base.model.NeuroBucketClassifier

data = [
    {"company": "ADIDAS", "category": "Shoes", "listPrice": 2000},
    {"company": "ADIDAS", "category": "Bracelets", "listPrice": 500},
    {"company": "NIKE", "category": "Shoes", "listPrice": 300},
    {"company": "NIKE", "category": "Clothes", "listPrice": 100},
    {"company": "PUMA", "category": "Clothes", "listPrice": 50},
    {"company": "PUMA", "category": "Shoes", "listPrice": 25},
    {"company": "PUMA", "category": "Shoes", "listPrice": 12},
    {"company": "NIKE", "category": "Shoes", "listPrice": 6},
    {"company": "NIKE", "category": "Shoes", "listPrice": 3},
]

model = train_from_dictionary(data, ["company", "category"], ["listPrice"])
print(model.get("listPrice", 1, '>')) # 3
print(model.get("listPrice", 4, '<')) # 2000

Note

The keyword_feats_name and bucket_feats_name should be contained in the data for each item

Module contents