Online sellers are already able to predict in advance around 50 percent of all purchase decisions made online – that is the result of the 9th ACM Recsys Challenge 2015 which was organized this year by the Yoochoose GmbH.
Over 500 teams participated in the prestigious competition well-known in the eCommerce and research industries. The teams from Russia, China and Canada took the three winning spots. They were best able to predict whether or not users would buy something on an eCommerce platform, as well as what they would buy, based on available data and algorithms they developed on their own. The conference and award presentation will take place between September 16 and 20 in Vienna. There are still some tickets available for interested visitors.
“The Recsys 2015 Challenge has shown that eCommerce today is better able than ever to adjust to the needs of its customers through recommender systems and thereby offer a personalized customer experience”, says Michael Friedmann, CEO of Yoochoose GmbH. “We are very happy that we were able to gain insightful data thanks to the help of several thousand participants, data which will help advance both future research and practical applications in the continually optimized use of Big Data in eCommerce. That is the only way for online sellers to increase customer retention and conversion rates in the future”.
The challenge posed to the teams from 49 countries participating in Recsys 2015 was to calculate whether or not a customer would buy something upon visiting a shop, as well as what he would buy, based on the dataset of an online seller containing information on user behavior up until the present. The victorious teams were able to correctly predict almost one in two customer decisions on the basis of their developed algorithms and methods, and will receive prizes worth a total of €5,000 provided by organizer Yoochoose.
The best solution to the challenge was presented by the Yandex Data Factory team, a subsidiary of the internationally active search engine operator from Russia, which placed ahead of a team from NetEase Youadao China & Tsinghua University and another team from the University of Toronto in the competition. Ratings were distributed on the basis of a point system. Aside from the accuracy of the prediction, the detailed explanation of the solution also influenced the final score. The papers will be published and available for download at the ACM Library following the conference.
Under the aegis of Yoochoose, the results are presented between September 16 and 20 on the Recsys Conference in Vienna to an audience of experts and discussed in detail in the course of a workshop. Users can expect important impulses for the use of systems in practice as well as insights on which submitted solutions work best. Interested parties from the eCommerce and research spaces can still register for a €100 fee.
Source: Yoochoose GmbH
