Print

Print


Hi Everyone, 

 

On Friday the 30th of September we will have a Statistics seminar at 2pm (room TBA). Our speaker will be Rajen Shah from the University of Cambridge. Please see below for the abstract. 

 

Before the talk we will have a lunch in the common room at 1pm. Please complete the form (here) by 4pm on Monday the 19th of September here if you’d like to attend the lunch. 

 

Please also let me know if you have speakers you’d like to invite. 

 

All the best, 

Henry 


 

 

Speaker: Rajen Shah 

Time: 2pm on Friday the 30th of September 

Title: The Projected Covariance Measure for assumption-lean variable significance testing 

Abstract: Testing the significance of a variable or group of variables X for predicting a response Y given additional covariates Z, is a ubiquitous task in statistics. A simple but common approach is to specify a linear model, and then test whether the regression coefficient for X is non-zero. However, when the model is misspecified, as will invariably be the case, the test may have poor power, for example when X is involved in complex interactions, or lead to many false rejections. 

In this work we study the problem of testing the model-free null of conditional mean independence, i.e. that the conditional mean of Y given X and Z does not depend on X. We propose a simple and general framework that can leverage flexible nonparametric or machine learning methods, such as additive models or boosted trees, to yield both robust error control and high power. The procedure involves using these methods to perform regressions, first to estimate a form of projection of Y on X and Z using one half of the data, and then to estimate the expected conditional covariance between this projection and Y on the remaining half of the data. While the approach is general, we show that a version of our procedure using spline regression achieves what we show is the minimax optimal rate in this nonparametric testing problem. 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Payables Team Mailbox [log in to unmask] 

 

Email – seminar 

 

 

Lunch 

 

NEED > a week notice!!! 

 

Microsoft Form – send to all Stats… 

[log in to unmask] 

Time  

Place 

 

Reimbursement form for non-staff members University of Bristol 

 

Budget code.. 

 

 

 

 

 

3 emails in a batch… 

 

 

In TB1 online… 

 

Prompt at the beginning… 

 

Friday afternoon slots: 

3pm – 5pm 

 

Previously offline: 

2pm-4pm 

 

45 minutes long talk + 5mins Q&A 

 

 

Starts Oct 1st 

 

Questions for Song: 

 

1.     Fortnightly 

 

2.     How to edit web-page 

 

3.     Email speakers – how far in advance 

 

A month and a half in advance 

 

Earlier for more senior people 

 

Two consecutive dates 

 

 

Always respond within a week… 

 

4.     Choice of areas? 

 

5.     Tea + Coffee for return to office 

 

6.     Travel costs 

 

7.     Speaker out to dinner/drink..? 

 

8.     Scheduling tool? 




To unsubscribe from the STATS-SEM list, click the following link:
https://www.jiscmail.ac.uk/cgi-bin/WA-JISC.exe?SUBED1=STATS-SEM&A=1