[Kaggle Series] Finding The "Missing" Missing Values

Missing value is usually considered as a boring and less important topic in data science. It is also a less discussed issue in many literatures. In this talk, Shu-Ting will go through all the fundamental theories of missing values and introduce us how to handle missing values, including classical and modern approaches, in a machine learning project. He will also give us a very interesting example of how he won the Silver Medal award in the historically largest Kaggle competition (8700+ teams) by "creating dummy missing values" in a dataset. Yes, missing values are not always bad. Sometimes creating missing values could significantly improve your model performance! Let’s look forward to Shu-Ting’s speech on how he found the "missing" missing values !

About the Speaker: Shu-Ting is an Applied Scientist at Amazon. He got his PhD in computational physics from UC Davis in 2015. After that, he completed his postdoc training in UC Irvine and UC Davis respectively. His is particularly interested in natural language processing, time series and semi-supervised learning.

Schedule:
6:30 - 7:00: "Doors" are open. Networking
7:00 - 8:00: Talk
8:00 - 8:15: Q&A from the audience
8:15 - 8:30: Networking

This event is hosted by DataCan to provide support to the 2022 Women in Data Science Datathon (2022 Jan to Feb): https://www.linkedin.com/feed/update/urn:li:activity:6880632915642974208

Join our community:
Slack: https://datacan.network/join-slack
Linkedin: https://www.linkedin.com/company/wids-vancouver/
Website: https://datacan.network/