Interpolation is a core numerical analysis technique used to estimate values between known data points, with widespread use in data processing, image processing, signal processing, and geographic information systems. Below is a detailed breakdown of common interpolation implementations in Python, alongside practical code examples.

Linear Interpolation

Linear interpolation assumes a linear relationship between adjacent known data points. In Python, the scipy.interpolate.interp1d function simplifies this workflow.

import numpy as np
from scipy.interpolate import interp1d

# Define known reference points
reference_x = np.array([1, 2, 3, 4])
reference_y = np.array([10, 20, 30, 40])

# Initialize linear interpolation function
linear_interp = interp1d(reference_x, reference_y)

# Generate new query points and compute interpolated values
query_points = np.array([1.5, 2.5, 3.5])
interp_results = linear_interp(query_points)

print("Linear interpolation results:", interp_results)

Polynomial Interpolation

Polynomial interpolation constructs a single polynomial function that passes through all given data points. Use numpy.polyfit to calculate polynomial coefficients, then numpy.poly1d to create callable interpolation functions.

import numpy as np

# Known input and output datasets
input_points = np.array([1, 2, 3, 4])
output_points = np.array([10, 20, 30, 40])

# Fit a degree n-1 polynomial for n reference points
poly_coefficients = np.polyfit(input_points, output_points, deg=len(input_points)-1)
poly_interp_func = np.poly1d(poly_coefficients)

# Run interpolation on new points
query_points = np.array([1.5, 2.5, 3.5])
poly_interp_results = poly_interp_func(query_points)

print("Polynomial interpolation results:", poly_interp_results)

Spline Interpolation

Spline interpolation uses piecewise low-order polynomials to create a smoother interpolated surface compared to high-degree polynomial interpolation. The scipy.interpolate.CubicSpline class implements cubic spline interpolation, the most common spline variant.

from scipy.interpolate import CubicSpline

# Reference data points
ref_x = np.array([1, 2, 3, 4])
ref_y = np.array([10, 20, 30, 40])

# Create cubic spline interpolation function
cubic_spline = CubicSpline(ref_x, ref_y)

# Compute interpolated values
query_points = np.array([1.5, 2.5, 3.5])
spline_interp_results = cubic_spline(query_points)

print("Cubic spline interpolation results:", spline_interp_results)

Common Interpolation Use Cases

1. Geographic Information Systems (GIS) Interpolation is used to estimate continuous spatial data from discrete sampling points, such as ground elevation or soil moisture levels.

   # Example: Interpolate elevation data across a geographic grid
   # (Using sample coordinate arrays for demonstration)
   sample_latitudes = np.array([29.9, 30.1, 30.3, 30.5])
   sample_longitudes = np.array([-90.1, -90.0, -89.9, -89.8])
   sample_elevations = np.array([100, 120, 140, 160])
   elevation_interp = CubicSpline(sample_latitudes, sample_elevations)
   interpolated_elev = elevation_interp([30.0, 30.2])
   

2. Image Processing Interpolation powers common image operations like resizing, rotation, and warping. OpenCV provides built-in interpolation methods for this workflow.

   import cv2
   # Resize an image with bilinear interpolation
   original_image = cv2.imread("sample_image.jpg")
   resized_image = cv2.resize(original_image, (1280, 720), interpolation=cv2.INTER_LINEAR)
   

3. Signal Processing Discrete sampled signals can be interpolated to increase sampling density for filtering, spectral analysis, or signal reconstruction.

   # Interpolate a sampled audio signal
   sample_time = np.array([0, 0.1, 0.2, 0.3])
   sample_signal = np.array([0, 0.5, 0.8, 0.3])
   signal_interp = interp1d(sample_time, sample_signal)
   high_res_time = np.linspace(0, 0.3, 100)
   high_res_signal = signal_interp(high_res_time)
   

4. Numerical Analysis Interpolation is used to approximate function values at unmeasured points for numerical approximation and curve fitting.

   # Approximate a mathematical function at sparse points
   func_x = np.array([0, np.pi/4, np.pi/2, 3*np.pi/4])
   func_y = np.sin(func_x)
   func_interp = CubicSpline(func_x, func_y)
   approx_value = func_interp(np.pi/3)
   

5. Financial Modeling Interpolation is used to build yield curves, fill missing stock price data, and price derivative financial instruments.

   # Build a simple interest rate curve
   tenor = np.array([0.5, 1, 2, 5])
   interest_rates = np.array([0.01, 0.012, 0.015, 0.02])
   rate_interp = interp1d(tenor, interest_rates)
   ten_year_rate = rate_interp(3)
   

6. Machine Learning & Data Mining Interpolation is used for data preprocessing, such as filling missing values in tabular datasets or smoothing noisy input features.

   import pandas as pd
   # Fill missing values in a dataset with linear interpolation
   raw_data = pd.DataFrame({"value": [10, None, 20, None, 30]})
   cleaned_data = raw_data.interpolate(method="linear")