Python二维数组基准归一化（Base Norming）实现方法咨询

Got it, let's work through this 2D array normalization problem together. You already nailed the 1D version, so extending that logic to 2D just needs a bit of row-wise handling to keep the first row and column intact.

Let's Break Down the Requirements #

• Keep the first row (headers) and first column (country names) exactly as they are.
• For each data row (starting from the second row):
  • The second element (2019 value) becomes 100 (our baseline).
  • Every subsequent element in the row is converted to a percentage relative to that row's 2019 value (i.e., (current_value / row_base) * 100), rounded to an integer to match your sample output.

Full Implementation Code #

Here's a straightforward way to achieve this:

# Original 2D array
lista = [['countries', 2019, 2021, 2022],['aruba', 2,13,8],['barbados', 6,34,39],['japan', 12,8,16]]

# Initialize the normalized array with the unchanged header row
listaNormed = [lista[0]]

# Process each data row starting from the second row
for row in lista[1:]:
    # Extract the country name and the base value (2019 data for this country)
    country_name = row[0]
    base_value = row[1]
    
    # Start building the normalized row: country name + baseline 100
    normalized_row = [country_name, 100]
    
    # Calculate normalized values for the rest of the row's elements
    for value in row[2:]:
        # Compute the percentage and convert to integer (matches your sample rounding)
        normalized_value = int((value / base_value) * 100)
        normalized_row.append(normalized_value)
    
    # Add the processed row to our result array
    listaNormed.append(normalized_row)

# Print the final normalized array
print(listaNormed)

What This Code Does #

1. Preserve Headers: We start by adding the original first row directly to listaNormed since we don't want to modify it.
2. Row-by-Row Processing: For each country row:
   • We grab the country name and its 2019 value (our baseline for that row).
   • We initialize the new row with the country name and the baseline value 100.
   • For each subsequent year's value, we calculate how it compares to the 2019 baseline, scale to a percentage, and convert to an integer (to match your sample output's whole numbers).
3. Build the Result: Each processed row gets added to listaNormed.

Running This Code Will Output Exactly What You Want: #

[['countries', 2019, 2021, 2022], ['aruba', 100, 650, 400], ['barbados', 100, 566, 650], ['japan', 100, 66, 133]]

Quick Note on Rounding #

If you wanted more precise decimal values instead of integers, you could replace int() with round() (e.g., round((value / base_value) * 100)). For your sample inputs, this would still give the same whole numbers, but it handles cases where you might need fractional percentages.

内容的提问来源于stack exchange，提问作者LabCoat