MASS: Product Kernel Density Visualization (KDV) Package

This package provides a C++ implementation of a Product Kernel Density Visualization (KDV) algorithm.

Input Data Format

The input data file should be plain text:

n
x1 y1
x2 y2
...
xn yn

• n = number of data points
• x and y in meters (or consistent units)

Input Parameters (for Python wrapper or command line)

• dim : data dimensionality (default: 2)
• method : algorithm method (0:SCAN, 1:SLAM, 2:MASS_CR, 3:MASS_OPT, 4:RQS_kd, 5:RQS_range)
• n_x, n_y : number of discrete regions along x/y-axis
• k_type_x, k_type_y : kernel type for x/y-axis (1:Epanechnikov, 2:Triangular, 3:Uniform)
• b_x_ratio, b_y_ratio : bandwidth ratio for x/y-axis kernel

Usage

Python Wrapper

from mass_pkdv import run_mass

run_mass(input_file, output_file)

Optionally, pass custom parameters:

run_mass(input_file, output_file, dim=2, method=2, n_x=800, k_type_x=3, b_x_ratio=0.8, n_y=600, k_type_y=3, b_y_ratio=0.8)

Command Line

bin\mass_pkdv.exe data\data.dat data\result.txt