Too many machine-learning-based data science solutions tend to concentrate their efforts on the specific stack used for learning, whether it’s the choice of big data framework or the precise machine-learning algorithm(s) used. The actual quality of the solution hinges on a good conceptual model of the domain, meaning that the outcomes optimized are much more important than the optimization technique. Moreover, the outcomes that need to be optimized are usually not the obvious ones.
A thorough understanding of the problem domain is essential to a good machine-learning-based product. Ofer Ron examines the development of one such product at LivePerson. LivePerson is the market leader in the ecommerce chat domain. One of the key challenges in this domain is targeting the consumers most likely to buy while best utilizing the chat agents, which are a scarce and expensive resource. Ofer shares LivePerson’s solution to this traffic-targeting problem, explaining the initial generic solution deployed, the failure of this solution, the analysis of the domain following this failure, and the eventual solution deployed with good results—as well as the software architecture necessary for implementing the solution.
Ofer Ron is a principal data scientist and architect at LivePerson, where he works on conceptualizing data products, researching them, and getting them deployed and running in production.
©2016, O'Reilly Media, Inc. • (800) 889-8969 or (707) 827-7019 • Monday-Friday 7:30am-5pm PT • All trademarks and registered trademarks appearing on oreilly.com are the property of their respective owners. • firstname.lastname@example.org
Apache Hadoop, Hadoop, Apache Spark, Spark, and Apache are either registered trademarks or trademarks of the Apache Software Foundation in the United States and/or other countries, and are used with permission. The Apache Software Foundation has no affiliation with and does not endorse, or review the materials provided at this event, which is managed by O'Reilly Media and/or Cloudera.