Normalization (pandas)

01 Jan 2024 in Study / Computer science on Pandas

pandas

• If data range of a specific column is so wide, it can have influence on the result of machine learning more than other columns.
• We should make the data range from each column be 0 ~ 1 or -1 ~ 1

Method

• 1: (a datum) / (maximum data of the column)
  • df[column’s name] = df[column’s name] / df[column’s name].max()
• 2: (a datum - minimum data) / (maximum data - minimum data)
  • df[column’s name] = (df[column’s name] - df[column’s name].min()) / (df[column’s name].max() - df[column’s name].min())