distances.py

#!/usr/bin/env python
#-*- coding:utf-8 -*-

from math import sqrt, fabs

"""
All distances used in the experiment
"""

__author__ =  'Filipe Wanderley Lima (fwl), Juliana Medeiros de Lucena (jml) and Tiago Ferreira Lima (tfl2)'

def euclidian_distance(a, b, solution):
    """
    Calculates Euclidian Distance of two vectors with numerical attributes

    @type a: list
    @param a: The vector
    @type b: list
    @param b: The other vector

    @rtype: float
    @returns: The Euclidian Distance between a and b
    """

    s = 0.0

    # Tirando label
    for i, ai in enumerate(a[:-1]):
        if solution[1][i]:
            wx = solution[0][i]
            s = s + (fabs(ai*wx - b[i]*wx)**2)

    return sqrt(s)

def manhattan_distance(a, b, solution):
    """
    Calculates Manhattan Distance of two vectors with numerical attributes

    @type a: list
    @param a: The vector
    @type b: list
    @param b: The other vector

    @rtype: float
    @returns: The Manhattan Distance between a and b
    """

    s = 0.0

    # Tirando label
    for ai in a[:-1]:
        i = a.index(ai)
        if solution[1][i]:
            wx = solution[0][i]
            s = s + fabs(ai*wx - b[i]*wx)

    return s

def adaptative_distance(a, b, distance_metric, training_set, solution):
    """
    Calculates Adaptative Distance of two vectors with numerical attributes

    @type a: list
    @param a: The vector
    @type b: list
    @param b: The other vector
    @type distance_metric: function
    @param distance_metric: The function used to calculate the distances

    @rtype: float
    @returns: The Adaptative Distance between a and b
    """

    return distance_metric(a, b, solution) / __min_sphere_radius(b, training_set, distance_metric, solution)

def __min_sphere_radius(a, training_set, distance_metric, solution):
    """
    Calculates the minimun distance between

    @type a: list
    @param a: The vector to center the sphere
    @type dataset: dict
    @param dataset: The dataset with all instances
    @type distance_metric: function
    @param distance_metric: The function used to calculate the distances

    @rtype: float
    @returns: The minimun radius sphere centered in a that ...
    """

    e = 0.01 #e > 0 is an arbitrarily small number

    distances = list()

    for c, l in training_set.items():
        if c != a[-1]:
            for li in l:
                distances.append(distance_metric(a, li, solution))

    distances.sort()

    if len(distances) != 0:
        return distances[0] - e
    else:
        return e