So many of our favorite daily activities are mediated by proprietary search algorithms. Whether you're trying to find a stream of that reality TV show on cat herding or shopping an eCommerce site for a new set of Japanese sushi knives, the relevance of search results is often responsible for your (un)happiness. Currently, small online businesses have no good way of evaluating the performance of their search algorithms, making it difficult for them to provide an exceptional customer experience.
The goal of this competition is to create an open-source model that can be used to measure the relevance of search results. In doing so, you'll be helping enable small business owners to match the experience provided by more resource rich competitors. It will also provide more established businesses a model to test against. Given the queries and resulting product descriptions from leading eCommerce sites, this competition asks you to evaluate the accuracy of their search algorithms.
Make a first submission with this
[Python benchmark](https://www.kaggle.com/users/993/ben-hamner/crowdflower-search-relevance/python-benchmark)
on
[Kaggle scripts](https://www.kaggle.com/c/crowdflower-search-relevance/scripts)
.
![](https://storage.googleapis.com/kaggle-media/competitions/kaggle/4407/media/cfbanner.png)
The dataset for this competition was created using query-result pairings enriched on the
[CrowdFlower](http://www.crowdflower.com/)
platform. They are sponsoring this competition as an investment in the open-source data science community. A dataset collected, cleaned, and labeled by CrowdFlower can make your supervised machine learning dreams come true.