Welcome to ltr++’s documentation!
Learning to Rank (LTR) is one of the methods that can be effectively applied to solve the task of creating a ranking model in Information Retrieval. It helps solving IR problems such as document retrieval, collaborative filtering, sentiment analysis, computational advertising etc. LTR method aims at learning a model that given a query and a set of candidate documents finds the appropriate ranking of documents according to their relevancy. LTR method aims at learning a model that given a query and a set of candidate documents finds the appropriate ranking of documents according to their relevancy. LTR++ is a Learning to Rank open-source library based on RankLib project. Currently two popular algorithms have been implemented and it also implements many retrieval metrics.
The LTR++ project is in early steps of development. Feel free to contribute or suggest features!