evaluation.py

import math
import numpy as np


def precision(actual, predicted, N):
    if isinstance(N, int):
        inter_set = set(actual) & set(predicted[:N])
        return float(len(inter_set))/float(N)
    elif isinstance(N, list):
        return np.array([precision(actual, predicted, n) for n in N])


def recall(actual, predicted, N):
    if isinstance(N, int):
        inter_set = set(actual) & set(predicted[:N])
        return float(len(inter_set))/float(len(set(actual)))
    elif isinstance(N, list):
        return np.array([recall(actual, predicted, n) for n in N])


def nDCG(Tr, topK, num=None):
    if num is None:
        num = len(topK)
    dcg, vec = 0, []
    for i in xrange(num):
        if topK[i] in Tr:
            dcg += 1/math.log(i+2, 2)
            vec.append(1)
        else:
            vec.append(0)
    vec.sort(reverse=True)
    idcg = sum([vec[i]/math.log(i+2, 2) for i in xrange(num)])
    if idcg > 0:
        return dcg/idcg
    else:
        return idcg


def mean_average_precision():
    pass


def AUC():
    pass