Build a Movie Recommendation System with Pearson Correlation in Python

Process

Use a web crawler to fetch Douban movie user information, define multi‑level rating categories, compute Pearson correlation between users, and analyze similarity either from a user‑centric or item‑centric perspective.

Movie rating classification

Very Bad

Bad

Average

Recommended

Strongly Recommended

Code

Data entry and storage:

# -*- coding: utf-8 -*-
import json
import sys
reload(sys)
sys.setdefaultencoding("utf-8")
user_info = {}
# Sample user data
user_dict = {
    'ns2250225':[4,3,4,5,4],
    'justin':[3,4,3,4,2],
    'totox':[2,3,5,1,4],
    'fabrice':[4,1,3,4,5],
    'doreen':[3,4,2,5,3]
}
def user_data(user_dict):
    for name in user_dict:
        user_info[name] = {}
        user_info[name][u'消失的爱人'] = user_dict[name][0]
        user_info[name][u'霍比特人3'] = user_dict[name][1]
        user_info[name][u'神去村'] = user_dict[name][2]
        user_info[name][u'泰坦尼克号'] = user_dict[name][3]
        user_info[name][u'这个杀手不太冷'] = user_dict[name][4]
user_data(user_dict)
# Save to file
try:
    with open('user_data.txt','w') as data:
        for key in user_info:
            data.write(key)
            for key2 in user_info[key]:
                data.write('\t')
                data.write(key2)
                data.write('\t')
                data.write(str(user_info[key][key2]))
                data.write('
')
except IOError as err:
    print('File error: ' + str(err))

Compute Pearson correlation and find similar users

from math import sqrt
def sim_pearson(prefs, p1, p2):
    si = {}
    for item in prefs[p1]:
        if item in prefs[p2]:
            si[item] = 1
    if len(si) == 0:
        return 0
    n = len(si)
    sum1 = sum([prefs[p1][it] for it in si])
    sum2 = sum([prefs[p2][it] for it in si])
    sum1Sq = sum([pow(prefs[p1][it], 2) for it in si])
    sum2Sq = sum([pow(prefs[p2][it], 2) for it in si])
    pSum = sum([prefs[p1][it] * prefs[p2][it] for it in si])
    num = pSum - (sum1 * sum2 / n)
    den = sqrt((sum1Sq - pow(sum1, 2) / n) * (sum2Sq - pow(sum2, 2) / n))
    if den == 0:
        return 0
    return num / den
# Insert my own data
user_info['me'] = {
    u'消失的爱人':5,
    u'神去村':3,
    u'炸裂鼓手':5
}
for user in user_info:
    res = sim_pearson(user_info, 'me', user)
    if res > 0:
        print('the user like %s is : %s' % ('me', user))
        print('result :%f' % res)

Recommend movies (weighted average)

def getRecommendations(prefs, person, similarity=sim_pearson):
    totals = {}
    simSums = {}
    for other in prefs:
        if other == person:
            continue
        sim = similarity(prefs, person, other)
        if sim <= 0:
            continue
        for item in prefs[other]:
            if item not in prefs[person] or prefs[person][item] == 0:
                totals.setdefault(item, 0)
                totals[item] += prefs[other][item] * sim
                simSums.setdefault(item, 0)
                simSums[item] += sim
    rankings = [(total / simSums[item], item) for item, total in totals.items()]
    rankings.sort()
    rankings.reverse()
    return rankings
res = getRecommendations(user_info, "me")
print('Recommend watching the movie:')
print(json.dumps(res, ensure_ascii=False))

Results and analysis

Users with similar taste: doreen, fabrice

Recommended movies for me: Titanic, Léon: The Professional

From a user‑centric view, finding like‑minded people helps discover potential interests.

From an item‑centric view, finding similar items helps identify potential customers.

Source: Data Mining: Introduction and Practice