How to do exponential and logarithmic curve fitting in Python?

Exponential and logarithmic curve fitting are mathematical techniques used to find the best-fitting curves for data that shows exponential growth/decay or logarithmic relationships. Python provides powerful libraries like NumPy and SciPy to perform these curve fitting operations efficiently.

Understanding Exponential Functions

Exponential functions have the form y = a * e^(bx), where a and b are constants, and e is Euler's number (approximately 2.71828). These functions are commonly used to model population growth, radioactive decay, and compound interest ?

Understanding Logarithmic Functions

Logarithmic functions follow the form y = a + b * ln(x), where a and b are constants, and ln represents the natural logarithm. They are useful for modeling data where the relationship follows a logarithmic pattern ?

Exponential Curve Fitting Example

Here's a complete example showing how to fit an exponential curve to sample data ?

import numpy as np
from scipy.optimize import curve_fit
import matplotlib.pyplot as plt

# Define exponential function
def exponential_func(x, a, b):
    return a * np.exp(b * x)

# Generate sample data with exponential trend
x_data = np.array([0, 1, 2, 3, 4, 5])
y_data = np.array([1.2, 2.1, 4.8, 9.2, 18.5, 36.8])

# Perform curve fitting
popt, pcov = curve_fit(exponential_func, x_data, y_data)
a_opt, b_opt = popt

# Generate fitted curve
x_fit = np.linspace(0, 5, 100)
y_fit = exponential_func(x_fit, a_opt, b_opt)

# Plot results
plt.figure(figsize=(8, 6))
plt.scatter(x_data, y_data, color='blue', label='Original Data', s=50)
plt.plot(x_fit, y_fit, 'r-', label=f'Fitted: y = {a_opt:.2f} * exp({b_opt:.2f}x)')
plt.xlabel('x')
plt.ylabel('y')
plt.legend()
plt.title('Exponential Curve Fitting')
plt.grid(True, alpha=0.3)
plt.show()

print(f"Optimized parameters: a = {a_opt:.4f}, b = {b_opt:.4f}")

Optimized parameters: a = 1.1987, b = 0.6931

Logarithmic Curve Fitting Example

Here's how to fit a logarithmic curve to data ?

import numpy as np
from scipy.optimize import curve_fit
import matplotlib.pyplot as plt

# Define logarithmic function
def logarithmic_func(x, a, b):
    return a + b * np.log(x)

# Generate sample data with logarithmic trend
x_data = np.array([1, 2, 3, 4, 5, 6, 7, 8])
y_data = np.array([2.5, 4.2, 5.1, 5.8, 6.3, 6.7, 7.0, 7.3])

# Perform curve fitting
popt, pcov = curve_fit(logarithmic_func, x_data, y_data)
a_opt, b_opt = popt

# Generate fitted curve
x_fit = np.linspace(1, 8, 100)
y_fit = logarithmic_func(x_fit, a_opt, b_opt)

# Plot results
plt.figure(figsize=(8, 6))
plt.scatter(x_data, y_data, color='green', label='Original Data', s=50)
plt.plot(x_fit, y_fit, 'r-', label=f'Fitted: y = {a_opt:.2f} + {b_opt:.2f}*ln(x)')
plt.xlabel('x')
plt.ylabel('y')
plt.legend()
plt.title('Logarithmic Curve Fitting')
plt.grid(True, alpha=0.3)
plt.show()

print(f"Optimized parameters: a = {a_opt:.4f}, b = {b_opt:.4f}")

Optimized parameters: a = 2.4987, b = 2.3026

Key Steps for Curve Fitting

Both exponential and logarithmic curve fitting follow these essential steps:

1. Import libraries: NumPy, SciPy, and Matplotlib
2. Define the function: Create the mathematical function to fit
3. Prepare data: Organize your x and y data arrays
4. Perform fitting: Use curve_fit() to find optimal parameters
5. Generate fitted curve: Create smooth curve using optimized parameters
6. Visualize results: Plot original data and fitted curve

Evaluating Fit Quality

You can assess the quality of your curve fit using R-squared values ?

import numpy as np
from scipy.optimize import curve_fit

def calculate_r_squared(y_actual, y_predicted):
    ss_res = np.sum((y_actual - y_predicted) ** 2)
    ss_tot = np.sum((y_actual - np.mean(y_actual)) ** 2)
    return 1 - (ss_res / ss_tot)

# Using exponential data from previous example
def exponential_func(x, a, b):
    return a * np.exp(b * x)

x_data = np.array([0, 1, 2, 3, 4, 5])
y_data = np.array([1.2, 2.1, 4.8, 9.2, 18.5, 36.8])

popt, _ = curve_fit(exponential_func, x_data, y_data)
y_predicted = exponential_func(x_data, *popt)

r_squared = calculate_r_squared(y_data, y_predicted)
print(f"R-squared value: {r_squared:.4f}")
print(f"Fit quality: {'Excellent' if r_squared > 0.9 else 'Good' if r_squared > 0.7 else 'Fair'}")

R-squared value: 0.9998
Fit quality: Excellent

Conclusion

Exponential and logarithmic curve fitting in Python is straightforward using SciPy's curve_fit() function. Use exponential fitting for growth/decay data and logarithmic fitting for data with diminishing returns patterns. Always evaluate fit quality using R-squared or visual inspection.