occ.py

from movielens.process import process_set
from utility import read_mapping
from sklearn import metrics
from random import shuffle
from statistics import mean, stdev

def occ(user, near_count, train_count, test_count, tree_count):
    users = read_mapping("./process/users.txt")
    movies = read_mapping("./process/movies.txt")
    near = read_mapping("./process/near.txt")[user][:near_count]
    train = process_set(user=user, count=train_count, users=users, movies=movies, mechanism=None)
    test = process_set(user=user, count=test_count, users=users, movies=movies, mechanism=None)

    trees = [list(near) for _ in range(tree_count)]
    for tree in trees:
        shuffle(tree)
    positives = [target[0] for target in train if target[1]]
    predictions = [distance_total(target[0], users, trees, positives) for target in test]
    targets = [1 if target[1] else 0 for target in test]
    return metrics.roc_auc_score(targets, predictions)

def distance_total(movie0, users, trees, positives):
    distance = 0
    for movie1 in positives:
        for tree in trees:
            distance += distance_single(movie0, movie1, users, tree)
    return distance / (len(trees) * len(positives))

def distance_single(movie0, movie1, users, tree):
    depth = 0
    for node in tree:
        if (movie0 in users[node]) == (movie1 in users):
            depth += 1
        else:
            return 0.5 ** depth
    return 0

USER = "Ariel-Garcé"
PATH = f"./experiment/occ-{USER}"

a = [occ(user=USER, near_count=5, train_count=100, test_count=100, tree_count=5) for _ in range(10)]
print(mean(a))
print(stdev(a))