1. def normalize_user_data(data, scaling_method='minmax'):
  2. """
  3. Normalizes user data.
  5. Args:
  6. data (list of dict): A list of dictionaries, where each dictionary
  7. represents user data.
  8. scaling_method (str): The scaling method to use. Options are 'minmax',
  9. 'zscore', or 'none'. Defaults to 'minmax'.
  11. Returns:
  12. list of dict: A list of dictionaries with normalized data. Returns
  13. an empty list if input is invalid.
  14. """
  16. if not isinstance(data, list):
  17. print("Error: Input data must be a list.")
  18. return []
  20. if not data:
  21. return []
  23. # Determine the keys (features) to normalize from the first user's data.
  24. keys = list(data[0].keys())
  26. if not keys:
  27. print("Error: No keys found in the data.")
  28. return []
  30. normalized_data = []
  32. for user in data:
  33. normalized_user = {}
  34. for key in keys:
  35. value = user.get(key)
  37. if value is None:
  38. normalized_user[key] = None #Handle missing values
  39. continue
  41. if scaling_method == 'minmax':
  42. min_val = min(user[key] for user in data)
  43. max_val = max(user[key] for user in data)
  44. if max_val == min_val:
  45. normalized_user[key] = 0.0 # Avoid division by zero
  46. else:
  47. normalized_user[key] = (value - min_val) / (max_val - min_val)
  48. elif scaling_method == 'zscore':
  49. mean = sum(user[key] for user in data) / len(data)
  50. std = (sum([(user[key] - mean)**2 for user in data]) / len(data))**0.5
  51. if std == 0:
  52. normalized_user[key] = 0.0 #Avoid division by zero
  53. else:
  54. normalized_user[key] = (value - mean) / std
  55. elif scaling_method == 'none':
  56. normalized_user[key] = value
  57. else:
  58. print(f"Warning: Unknown scaling method '{scaling_method}'. Skipping normalization for {key}")
  59. normalized_user[key] = value
  62. normalized_data.append(normalized_user)
  64. return normalized_data

Add your comment