We are thrilled to announce that LogicAI took the 1st place in the ACM RecSys 2019 competition. RecSys is a prestigious, yearly data science competition aimed at building recommendation systems, based on data provided by different companies every year. Our Team from LogicAI participated in RecSys for the first time and reached the top of the leaderboard.
This year, the competition was hosted by Trivago. The overview of the competition can be found here (https://recsys.trivago.cloud/challenge/). Trivago is an international hotel search platform, allowing travelers to choose and compare their accommodation easily, everywhere in the world. This year’s challenge was focused on travel metasearch. The goal of RecSys 2019 competition was to develop a session based and context-aware recommender system, using various input data, to provide a list of accommodations that will match the needs of the user. In this competition, our task was to predict which accommodations have been clicked in the search result during the last part of a user session in an offline evaluation setup.
Our Data Science Team in LogicAI has shown its experience and passion for data science in numerous competitions in the past (www.logicai.io). Except for providing value for our customers through realizing consulting projects from data science area, we also believe the best way to develop new skills in our Team, given the dynamic environment of the modern world, is through teamwork on hands-on tasks.
As creators of the largest data science offline community in the world, we believe that the best opportunity for Data Scientists to improve their skills and develop, is nowadays provided by online competitions, eg. on Kaggle (www.kaggle.com) platform. This is why we as LogicAI partnered up with Kaggle and brought to life the idea of Kaggle Days – offline events for data scientists to network, learn through practice and compete in teams for amazing prizes in day-long Data Science competitions.
In RecSys 2019 competition, our Team created a set of effective features that helped us boost state-of-the-art algorithm a lot. The goal was to provide a user with the best recommendations of accommodation, based on their needs. We are proud to say that with our diverse backgrounds, we managed to reach as high as 1st place on the leaderboard.
We are happy to have achieved this together and we will gladly try our skills in future endeavors of data science competitions.
Author other articles
3 languages for Machine Learning
Realtime recommendation systems for session-based user activities
How to build event-based models