Intelligent Recommendation

Based on advanced big data and artificial intelligence technology of JD, Intelligent Recommendation digs out perfect group packages and profile tags as per commodities launched by the merchant in the advertising scenario and guides the launch, to help the merchants increase their advertisement benefits.

Core Advantages

  • Individualization

    Change the traditional uniform operation mode, and realize the individual display and customized intelligent operation through data and models

  • Big Data Analysis

    With JD Cloud’s big data analysis capacity, users’ interest and preference data are overall and centrally analyzed, effectively solving problems such as cold start and data sparsity in the business course

  • Machine Learning

    Machine learning system of large-scale real-time computer, advanced algorithm technology and effect evaluation platform

  • Plug and Play

    Integrate with the customer’s system in the plug-in form, having no influence on existing system architecture of the customer, low development workload and short duration

Business Architecture

Typical Scenario

Individual Recommendation

Typical Scenario: The home page will provide you with commodity information streams such as the recommended, your favorite and similar commodity recommendation

Solution: Deeply dig out users’ interests, features and behavior data, adopt machine-learning model, calculate users’ commodity preference and realize individual commodity display on pages

Social Group Recommendation

Typical Scenario: Under the social group e-commerce scenario, recommend commodities for distribution WeChat groups and the Circle of Friends

Solution: Based on user behaviors and transaction data under the WeChat scenario, analyze consumer groups’ profile features, calculate consumer groups’ potential preferential commodities and recommend commodities by machine learning

Collocation Recommendation

Typical Scenario: Commodity recommendation on the order settlement page, commodity collection page, bundle sales and other scenarios

Solution: Based on transaction data of massive commodities, bundle sales association is analyzed with the algorithm model and bundle sales is recommended in combination with users’ behavior

Relevant Recommendation

Typical Scenario: Recommend related commodities on the commodity details and similar product scenarios

Solution: Based on machine-learning algorithm, calculate commodity similarity and recommended related commodity in combination with user’s behavior

Hot Recommendation

Typical Scenario: Hot recommendation on the home page, ranking list and other positions

Solution: Through multi-dimensional data analysis, hot commodities are pushed on the home page and commodity ranking is optimized according to a user’s interest, preference and behavior characteristics