Recommender System

This week, our teacher introduced us to a fascinating topic: Recommender Systems. Let's provide you with a brief overview of what a recommender system is and how it works.

A recommender system is a type of information filtering system that predicts and suggests items or content that a user might be interested in, based on their preferences, past behaviour, or similarities with other users. These systems are widely used in various online platforms such as e-commerce websites, streaming services, social media platforms, and more.

Here's how a recommender system typically works:

1. Data Collection: The system gathers data on user interactions, such as purchases, ratings, views, clicks, likes, and other relevant actions.

2. User Profiling: The system creates profiles for individual users based on their preferences, behaviours, and historical interactions with items or content.

3. Item Representation: Similarly, the system analyses items or content available in the system and represents them in a meaningful way, such as keywords, categories, or features.

4. Recommendation Generation:

   • Collaborative Filtering: This method identifies similar users or items based on their past interactions. For user-based collaborative filtering, recommendations are made by finding items that similar users have liked or interacted with. For item-based collaborative filtering, recommendations are made by finding items similar to those that the user has interacted with in the past.

   • Content-Based Filtering: This method recommends items or content similar to those that the user has previously liked or interacted with. It analyzes the attributes or features of items and compares them with the user's preferences.

   • Hybrid Methods: Some recommender systems combine collaborative filtering and content-based filtering approaches to provide more accurate and diverse recommendations.

5. Ranking and Presentation: The system ranks the recommended items based on their relevance to the user and presents them through various interfaces such as personalized suggestions, recommendation lists, or targeted advertisements.

6. Feedback Loop: The system continuously learns and improves its recommendations based on user feedback, including explicit feedback (e.g., ratings, reviews) and implicit feedback (e.g., clicks, dwell time).

Overall, recommender systems play a crucial role in enhancing user experience, increasing user engagement, and driving business growth by delivering personalized and relevant recommendations to users.

Putting Theory into Practice: Using Anaconda

We gained practical experience by employing Anaconda, a widely used Python tool for data science, to put into practice the concepts covered in the course. Initially, we imported essential libraries and proceeded to examine a dataset containing movie ratings. This dataset offered insights into user interactions with movies, detailing who rated which movies and when. Through analyzing this data, we gleaned information such as the number of ratings each user provided and the number of ratings each movie garnered. Such analysis enhanced our comprehension of the dataset and provided insights into the level of user engagement with the movies.

The head() shows first five rows from the top of the data set. We can change it to different numbers to see different rows. If I change it to 10 it will show first ten rows of the data set.

And tail() gives the first 5 rows from the bottom and we can put the numbers between the brackets to see different rows from bottom.

Movie rating based on the inputs n_ratings = len(ratings) n_movies = len(ratings['movieId'].unique()) n_users = len(ratings['userId'].unique()) print(f"Number of ratings: {n_ratings}") print(f"Number of unique movieId's: {n_movies}") print(f"Number of unique users: {n_users}") print(f"Average ratings per user: {round(n_ratings/n_users, 2)}") print(f"Average ratings per movie: {round(n_ratings/n_movies, 2)}")

user_freq = ratings[['userId', 'movieId']].groupby(

'userId').count().reset_index()

user_freq.columns = ['userId', 'n_ratings']

user_freq.head(10) #Now we display the first 5 rows of our newly created table "user_freq"

Comprehending the concept of a "cluster" holds significant importance, particularly in the realm of data science. In this context, a cluster refers to a collection of data points that are grouped together due to their similarity in certain aspects. Think of it as akin to a club where members share common characteristics, making them more akin to each other than individuals in other clubs or groups. Clusters play a crucial role in various machine learning applications, particularly in unsupervised learning scenarios. In unsupervised learning, the objective is to discern patterns in data without pre-existing labels. For instance, in our recommender system, clustering can be employed to categorize similar users or items based on their attributes or behaviors, such as movie preferences. This facilitates the system in providing more accurate recommendations by assuming that individuals within the same cluster are likely to have similar movie preferences.

The main part of this exercise was making a user-item matrix, which is essential for recommender systems. We used the SciPy library to create this matrix. It shows how users rate items like movies. This matrix is crucial for building personalized recommendations.