Розробка рекомендаційної системи для функціонування в умовах недостатньої визначеності уподобань користувачів

Abstract

In the work was developed recommendation system for operation in low certainty preferences of users. K-means and TF-IDF algorithms was selected for implementation as a result of problem area analization. Was designed the architecture of the hybrid recommendation system and recommendation balanser that absorbed the benefits of content and collaborative approaches. This software implementation is an example of using artificial intelligence to solving classification problems. Recommendation system implementation was done with best practice of software development compliance. The system is scalable and secure with OAuth 2.0 protocol. Software implementation was involved with database MySQL, used programming language Java 7 in the programming environment IntelliJ IDEA 13.0. Passed functional and load testing using JUnit 4 library.