# r-squared value calculation
[docs]
def r2(y_pred, y_true):
"""
Calculates r-squared using linear regression.
Computes the r-squared value (coefficient of determination) using the provided y_pred
list and y_true list.
Parameters
----------
y_pred : list
y_pred values to be used in calculating r-sqaured value.
y_true : list
y_true values to be used in calculating r-sqaured value.
Returns
-------
float
r-sqaured value which is <= 1. 1 is the best score and a score below 0 is worse than
using the mean of the target as predictions.
Examples
--------
>>> data = {
>>> 'math_test': [80, 85, 90, 95],
>>> 'science_test': [78, 82, 88, 92],
>>> 'final_grade': [84, 87, 91, 94],
>>> 'absences': [3, 0, 1, 30]
>>> }
>>> r2(data['math_test'],data['final_grade'])
0.997
>>> r2(data['math_test'],data['absences'])
0.541
"""
from sklearn.linear_model import LinearRegression
import numpy as np
# Ensure inputs are lists
if not isinstance(y_pred, list) or not isinstance(y_pred, list):
print('Input must be lists')
return None
# Check for empty lists
if len(y_pred) == 0 or len(y_true) == 0:
print('Input cannot be empty')
return None
# Convert lists to NumPy arrays if they are still in list format
if isinstance(y_pred,list):
y_pred = np.array(y_pred)
if isinstance(y_true,list):
y_true = np.array(y_true)
# Create a linear regression model and fit it to the data
model = LinearRegression()
model.fit(y_pred.reshape(-1,1),y_true)
y_true_predicted = model.predict(y_pred.reshape(-1,1))
# Calculate the mean of y_true
y_true_mean = np.mean(y_true)
# Compute Residual Sum of Squares
RSS = sum(((y_true-y_true_predicted) ** 2))
# Compute Total Sum of Squares (TSS)
TSS = sum(((y_true-y_true_mean) ** 2))
# Compute R-squared value and round it to 3 decimal places
return round(1 - (RSS/TSS),3)