Built independently by an author, for readers. Read the story and support ChapterPal
ChapterPal Logo
Log inSign up

keyword

dimensionality reduction techniques

Dimensionality reduction techniques are computational and statistical methods used in machine learning and data analysis to reduce the number of input variables or features in a dataset while preserving its essential information and underlying structure. When datasets contain large numbers of variables, high dimensionality often leads to increased computational complexity, noise, and data sparsity. These techniques address such challenges through feature selection, which isolates and retains the most informative original variables, and feature extraction, which transforms high-dimensional data into a lower-dimensional representation using mathematical transformations such as principal component analysis, singular value decomposition, or autoencoders. By compressing the feature space, dimensionality reduction enhances algorithmic efficiency, mitigates overfitting, aids data visualization, and improves the scalability and accuracy of predictive models.

1 item

Amazon.com recommendations: item-to-item collaborative filtering

Amazon.com recommendations: item-to-item collaborative filtering

Greg Linden, Brent Smith, Jeremy York

OrganizationsAmazon

Why you should read this

Demonstrates the architectural implementation of item-based filtering that allowed recommendations to scale to millions of items in production.

Part of:Must-read papers on recommender systems

Recommendation algorithms are best known for their use on e-commerce Web sites, where they use input about a customer's interests to generate a list of recommended items. Many applications use only the items that customers purchase and explicitly rate to represent their interests, but they can also use other attributes, including items viewed, demographic data, subject interests, and favorite artists. At Amazon.com, we use recommendation algorithms to personalize the online store for each customer. The store radically changes based on customer interests, showing programming titles to a software engineer and baby toys to a new mother. There are three common approaches to solving the recommendation problem: traditional collaborative filtering, cluster models, and search-based methods. Here, we compare these methods with our algorithm, which we call item-to-item collaborative filtering. Unlike traditional collaborative filtering, our algorithm's online computation scales independently of the number of customers and number of items in the product catalog. Our algorithm produces recommendations in real-time, scales to massive data sets, and generates high quality recommendations.

Added

2026-01-25