Based on its cutting-edge big data and artificial intelligence (AI) technologies and years of experience in the e-commerce industry, Alibaba provides Artificial Intelligence Recommendation for developers. Artificial Intelligence Recommendation is a personalized recommendation service that can help increase the customer purchase rate and order conversion rate.

Overview

This topic describes how to use Artificial Intelligence Recommendation to build an intelligent system for recommending items on e-commerce websites. The system uses the following services of Alibaba Cloud: Log Service for collecting logs, ApsaraDB for RDS as the back-end data service, MaxCompute as a data warehouse, and DataWorks for synchronizing and processing data.

Recommendation information on e-commerce websites includes product properties such as logistics information and sales information. Such information appears on homepages or as news feeds and can promote transactions and increase the customer purchase rate and order conversion rate.

Scenarios

  • Displays users' preferred and personalized contents in a waterfall flow view.
  • Displays recommendations related to a product on the product details page.
  • Displays focus pictures and popular recommendations on website homepages.
  • Displays news, live videos, and social events.

Process

The process of implementing intelligent recommendation on e-commerce websites is as follows:
  1. Log Service collects user behavior logs.
  2. Log Service uploads user behavior logs to MaxCompute.
  3. Log Service uploads product and user data to MaxCompute.
  4. MaxCompute pushes data to Artificial Intelligent Recommendation, which then uses AI algorithms to generate recommendations based on the data.

Highlights

  • Artificial Intelligent Recommendation is a mature solution developed by Alibaba based on its years of experience. It is suitable for various industries and scenarios and can meet diversified business needs of different users.
  • DataWorks in this solution frees you from concerns about stream processing, algorithms, O&M, and monitoring.
  • Data is isolated between users and sensitive information is encrypted, which guarantee information security.
  • Advanced algorithm technologies, multi-mode integration, efficient cold start, real-time policy adjustment, and model training are used, freeing you from manual operations.
  • Multiple services are seamlessly integrated, which enables data synchronization within hours.

Details

For more information, see Best practice of intelligently recommending items on e-commerce websites.