- python knowledge
- Base of data manipulation numpy, pandas, csv
- Base of Machine learning and Deep learning
Data I'm using to work here come from GoodRead and Amazon
The link will be here as soon as possible
""" Recommender systems are tools for interacting with large and complex information spaces. They provide a personalized view of such spaces, prioritizing items likely to be of interest to the user. The field, christened in 1995, has grown enormously in the variety of problems addressed and techniques employed, as well as in its practical applications. Recommender systems research has incorporated a wide variety of artificial intelligence techniques including machine learning, data mining, user modeling, case-based reasoning, and constraint satisfaction, among others. Personalized recommendations are an important part of many on-line e-commerce applications such as Amazon.com, Netflix, and Pandora. This wealth of practical application experience has provided inspiration to researchers to extend the reach of recommender systems into new and challenging areas. The purpose of this special issue is to take stock of the current landscape of recommender systems research and identify directions the field is now taking.
Recommender Systems are computer algorithms that analyse user behaviour and generate personalized recommendations for products, services, or content. These systems are widely used by various online platforms such as YouTube, Amazon, TikTok, etc to enhance user engagement and satisfaction. By leveraging machine learning and deep learning techniques, these systems can provide highly accurate and personalized recommendations to users, leading to increased user engagement and revenue for the platforms. For example, YouTube uses recommendation algorithms to suggest videos based on a user's viewing history and preferences (Friedland et al., 2010). Similarly, Amazon recommends products based on a user's purchase history and browsing behaviour (Linden et al., 2003). TikTok uses a combination of user engagement metrics and content analysis to generate personalized video feeds (Zhou et al., 2021). Recommender Systems have become increasingly popular in recent years due to the explosive growth of online platforms and the vast amounts of user-generated data that they generate. These systems are beneficial for both users and platforms, as users receive personalized recommendations and platforms can increase user engagement and revenue. Overall, Recommender Systems play an essential role in enhancing user experience and driving business growth for online platforms.
The primary function of a recommender system is to predict future user preferences based on historical data. These systems are now used in various aspects of our lives, such as purchasing items online, selecting movies, and adding friends on social media. However, predicting user preferences is a complex problem that has been the subject of intensive research for many years. One notable event that sparked increased interest in this area was the Netflix competition, which offered a 1-million-dollar prize to researchers who could substantially improve the quality of recommendations generated.
Two primary approaches to generating recommendations are used in recommender systems. One approach involves predicting the rating a user would give to a specific item in the system, while the other approach predicts a set of items presented as an ordered list that would be recommended to the user. Recommender systems can be divided into personalized and non-personalized systems. Non-personalized systems draw conclusions based on the global behaviour of all users in the system, while personalized systems create a user profile based on their historical activity to generate recommendations.
A recommender system typically consists of a set of users (U) and a set of items (I). The interactions between users and items are recorded in a matrix R, where each element (u,i,rui) represents a user u interacting with an item i and giving it a rating rui. Several techniques and methods have been proposed for recommender systems, including content-based filtering, collaborative filtering, knowledge-based filtering, and hybrid approaches.
""" from https://www.researchgate.net/publication/220604600_Recommender_Systems_An_Overview
- Memory based recommendation
- Model based recommendation
Memory-based recommendation systems rely on historical data to predict future user preferences and generate recommendations. There are three commonly used similarity metrics in memory-based recommendation systems: Jaccard, cosine, and Pearson. Each metric has its advantages and disadvantages, which can impact the quality and accuracy of recommendations.
Jaccard similarity measures the similarity between two sets by computing the ratio of the intersection of the sets to the union of the sets. The advantage of using Jaccard similarity is that it works well for sparse datasets, where there are many missing values. Jaccard similarity can also handle binary data, which is useful when dealing with user-item interactions (i.e., when a user either interacts with an item or does not). However, Jaccard similarity does not take into account the magnitude of the data and is sensitive to noise, which can result in less accurate recommendations. Jaccard similarity is commonly used in collaborative filtering systems.
Cosine similarity measures the similarity between two vectors by computing the cosine of the angle between them. The advantage of using cosine similarity is that it is computationally efficient and works well for high-dimensional data. Cosine similarity can handle both binary and non-binary data, making it a versatile similarity metric. However, cosine similarity does not take into account the magnitude of the data and can be biased towards popular items. Cosine similarity is commonly used in content-based recommendation systems.
Pearson correlation measures the linear correlation between two variables. The advantage of using Pearson correlation is that it takes into account the magnitude of the data and can handle both positive and negative correlations. Pearson correlation can also handle missing data, making it useful for datasets with many missing values. However, Pearson correlation assumes that the data is normally distributed, which may not always be the case. Pearson correlation is commonly used in collaborative filtering systems.
In general, the choice of similarity metric depends on the nature of the dataset and the recommendation system being developed. For example, Jaccard similarity is commonly used in collaborative filtering systems, while cosine similarity is commonly used in content-based recommendation systems. Pearson correlation can be used in both collaborative filtering and content-based recommendation systems. It is also common to use a combination of similarity metrics to improve the quality and accuracy of recommendations.
In contrast to memory-based recommendation approaches, model-based approaches seek to learn parameterized representations of users and items, so that recommendations can be made in terms of the learned parameters.