51 分鐘
Matt and Ulrik make unsupervised product recommendation engines Power BI & More
-
- 科技
This episode is brought to you by by Maplytics by Inogic.
Data Scientist Matt Lamb and Microsoft MVP Ulrik Carlsson discusses how you create product recommendation engines. A separate discipline in data science, combining content filtering and collaborative filtering, to do targeted product recommendations is not only more difficult, but possibly also one of the most lucrative.
Episode also includes in discussions on:
Combining advanced customer profiling with transactional data.
Matt talks to his new product PinPoint, a product recommendation engine for the Aftermarket How Content Filtering and Collaborative Filtering combined can make for advanced product recommendations Why Ulrik doesn't like continued recommendations from Amazon to buy smoke detectors when they perfectly well know he already has two (and how to tune your algorithm to avoid annoying your customer). Possible data science urban legend on Target identifying teenage pregnancies before concerned parents of pregnant teen knows about it. Will Matt this time give a concrete answer to the question on how many records are needed to get good results from these algorithms?
Links: PinPoint for Aftermarket
This episode is brought to you by by Maplytics by Inogic.
Data Scientist Matt Lamb and Microsoft MVP Ulrik Carlsson discusses how you create product recommendation engines. A separate discipline in data science, combining content filtering and collaborative filtering, to do targeted product recommendations is not only more difficult, but possibly also one of the most lucrative.
Episode also includes in discussions on:
Combining advanced customer profiling with transactional data.
Matt talks to his new product PinPoint, a product recommendation engine for the Aftermarket How Content Filtering and Collaborative Filtering combined can make for advanced product recommendations Why Ulrik doesn't like continued recommendations from Amazon to buy smoke detectors when they perfectly well know he already has two (and how to tune your algorithm to avoid annoying your customer). Possible data science urban legend on Target identifying teenage pregnancies before concerned parents of pregnant teen knows about it. Will Matt this time give a concrete answer to the question on how many records are needed to get good results from these algorithms?
Links: PinPoint for Aftermarket
51 分鐘