The National B2B e-Commerce Platform serves as a robust trading system for commercial companies across the nation. As part of this project, I played a key role in developing the advertising server utilizing Revive AdServer and implementing a RESTful service for click-through rate counting and recommendation. Additionally, I developed an on-site search engine based on Apache Solr, further enhancing the platform’s functionality.
With a focus on web service development, I utilized Flask, Revive AdServer, Apache Solr, and Redis to create a RESTful service that provided real-time click-through rate counting and high concurrency capabilities. The integration of Redis as a backend ensured efficient and reliable storage of statistical data.
To enhance user experience and improve product visibility, I developed a recommendation module using collaborative filtering and ranking algorithms. This module significantly improved the ranking correlation and achieved an impressive F1 score increase of 0.07 on the MovieLens dataset compared to traditional UserCF/ItemCF algorithms. The recommendation module played a pivotal role in optimizing the National B2B project’s performance.
Highlights of this project include the successful integration of Revive AdServer and Apache Solr search engine, the development of a high-concurrency RESTful service for click-through rate counting, and the implementation of a recommendation module utilizing collaborative filtering and ranking algorithms.
Throughout the project, my skills in Java, REST, Redis, ad server systems, and search engine technologies were instrumental in driving the development and success of the National B2B e-Commerce Platform.