Content Based Recommendations

Introduction and Context Setting

Hello and welcome! In today's lesson, we will dive deeper into content-based recommendation systems . Before we explore building recommendations for the dataset from the previous lesson, let's consider a simpler example. The aim here is to learn how to compute similarity and make recommendations effectively using this simpler context. By the end of this lesson, you will be empowered to understand the mechanics and apply them to different datasets.

Understanding Profiles

In a content-based recommendation system, there are two main profiles to understand: item profiles and user profiles. Item Profiles refer to the attributes or features of an item. For example, movies can be represented by their genres. In C++, we can use a struct to represent a movie profile and vectors to store the genre values. C++#include <vector> #include <string> // List of genres (for reference) const std::vector<std::string> GENRES = {"Action", "Romance", "Sci-Fi"}; // Struct to represent an item (movie) profile struct ItemProfile { std::string name; std::vector<double> genres; // genre values in the same order as GENRES }; // Sample item profiles: described by movie genres std::vector<ItemProfile> item_profiles = { {"Movie1", {0.9, 0.1, 0.5}}, {"Movie2", {0.2, 0.8, 0.1}}, {"Movie3", {0.4, 0.9, 0.2}} };#include <vector> #include <string> // List of genres (for reference) const std::vector<std::string> GENRES = {"Action", "Romance", "Sci-Fi"}; // Struct to represent an item (movie) profile struct ItemProfile { std::string name; std::vector<double> genres; // genre values in the same order as GENRES }; // Sample item profiles: described by movie genres std::vector<ItemProfile> item_profiles = { {"Movie1", {0.9, 0.1, 0.5}}, {"Movie2", {0.2, 0.8, 0.1}}, {"Movie3", {0.4, 0.9, 0.2}} }; In this code, each movie is represented by a vector of values, each corresponding to a genre. The values range between 0 and 1, indicating the strength of the association. User Profiles capture a user's preferences, often based on their past behavior. For example, a user may like "Action" and "Sci-Fi" movies. We can represent the user profile as a vector of values, in the same order as the genres. C++// Sample user profile: representing preferences for genres // User has watched and liked past movies mostly in 'Action' and 'Sci-Fi' std::vector<double> user_profile = {0.8, 0.1, 0.6};// Sample user profile: representing preferences for genres // User has watched and liked past movies mostly in 'Action' and 'Sci-Fi' std::vector<double> user_profile = {0.8, 0.1, 0.6}; The user profile is similar in structure to an item profile, allowing direct comparison of preferences to item attributes.

Computing Similarity

A critical part of recommending items is determining how similar they are to a user's preferences. We compute this similarity using the dot product, a straightforward mathematical operation ideal for this purpose. The dot product sums the products of corresponding features in the user and item profiles. Here's how it's implemented in C++: C++// Compute similarity using a dot product double compute_similarity(const std::vector<double>& user, const std::vector<double>& item) { double sum = 0.0; for (size_t i = 0; i < user.size(); ++i) { sum += user[i] * item[i]; } return sum; }// Compute similarity using a dot product double compute_similarity(const std::vector<double>& user, const std::vector<double>& item) { double sum = 0.0; for (size_t i = 0; i < user.size(); ++i) { sum += user[i] * item[i]; } return sum; } In this function, each feature's value in the user profile is multiplied by the corresponding value in the item profile, and these products are summed up. This sum gives us a similarity score, indicating how well an item matches the user's preferences. Let's see an example of how we can calculate the similarity score for Movie1 using the provided user profile: Action: 0.8 (User) * 0.9 (Movie1) = 0.72 Romance: 0.1 (User) * 0.1 (Movie1) = 0.01 Sci-Fi: 0.6 (User) * 0.5 (Movie1) = 0.3 The total similarity score for Movie1 is 0.72 + 0.01 + 0.3 = 1.03.

Join the 1M+ learners on CodeSignal

Be a part of our community of 1M+ users who develop and demonstrate their skills on CodeSignal

Making Recommendations

With similarity scores in hand, we can now make recommendations by selecting the items with the highest scores. Let's go through the recommendation process:

We will create a struct to store the recommendation results and then sort them manually.

C++

#include <iostream>
#include <vector>
#include <string>
#include <algorithm>

// Struct to store recommendation results
struct Recommendation {
    std::string name;
    double score;
};

// Function to recommend items based on similarity to user profile
std::vector<Recommendation> recommend_items(const std::vector<double>& user_profile, 
                                           const std::vector<ItemProfile>& items) {
    std::vector<Recommendation> recommendations;
    
    for (const auto& item : items) {
        Recommendation rec;
        rec.name = item.name;
        rec.score = compute_similarity(user_profile, item.genres);
        recommendations.push_back(rec);
    }

    // Sort recommendations by score in descending order
    std::sort(recommendations.begin(), recommendations.end(), 
              [](const Recommendation& a, const Recommendation& b) {
                  return a.score > b.score;
              });
    
    return recommendations;
}

int main() {
    std::vector<Recommendation> recommendations = recommend_items(user_profile, item_profiles);

    std::cout << "Recommendations:" << std::endl;
    for (const auto& rec : recommendations) {
        std::cout << rec.name << ": " << rec.score << std::endl;
    }

    return 0;
}

#include <iostream>
#include <vector>
#include <string>
#include <algorithm>

// Struct to store recommendation results
struct Recommendation {
    std::string name;
    double score;
};

// Function to recommend items based on similarity to user profile
std::vector<Recommendation> recommend_items(const std::vector<double>& user_profile, 
                                           const std::vector<ItemProfile>& items) {
    std::vector<Recommendation> recommendations;
    
    for (const auto& item : items) {
        Recommendation rec;
        rec.name = item.name;
        rec.score = compute_similarity(user_profile, item.genres);
        recommendations.push_back(rec);
    }

    // Sort recommendations by score in descending order
    std::sort(recommendations.begin(), recommendations.end(), 
              [](const Recommendation& a, const Recommendation& b) {
                  return a.score > b.score;
              });
    
    return recommendations;
}

int main() {
    std::vector<Recommendation> recommendations = recommend_items(user_profile, item_profiles);

    std::cout << "Recommendations:" << std::endl;
    for (const auto& rec : recommendations) {
        std::cout << rec.name << ": " << rec.score << std::endl;
    }

    return 0;
}

The function recommend_items calculates the similarity score for each item and stores the results. It then sorts the items by these scores in descending order to prioritize recommendations that best align with user interests.

Output:

Recommendations:
Movie1: 1.03
Movie3: 0.53
Movie2: 0.3

Recommendations:
Movie1: 1.03
Movie3: 0.53
Movie2: 0.3