לוח שנה

S M T W T F S
 
 
 
 
 
 
1
 
2
 
3
 
4
 
5
 
6
 
7
 
8
 
9
 
10
 
11
 
12
 
13
 
14
 
15
 
16
 
17
 
18
 
19
 
20
 
21
 
22
 
23
 
24
 
25
 
26
 
27
 
28
 
29
 
30
 
 
 
 
 
 
 

"In Conversation" Series | Ya’acov Ritov | Bayesian statistics without beliefs and partial linear models

Date: 
Sun, 02/06/202414:00
pras-isaerl-yaakov-ritov

"In Conversation" Series

 

Lecturer: 

Prof. Ya’acov Ritov (Hebrew University)

Title: 

Bayesian statistics without beliefs and partial linear models

Abstract: 

I'll survey two lines of 40 years of work. The first would deal with the possibility of Bayesian inference in the age of big data. I'll present examples where belief-based Bayesian statistical inference is impossible to achieve the basic criteria of the asymptotic frequentist principles. I'll discuss the compound decision and empirical Bayes approaches and argue for the existence of optimal procedures (work in progress). 

The partial linear model is a regression model in which the regression function is a sum of a linear component and a nonparametric one. I'll discuss its semiparametric roots and modern ramifications as the inference in ultra-high dimensional regression with more variables than subjects. I'll finish by describing the regression discontinuity model as a partial linear model. This model is used to analyze, for example, the effect of a selective program on its students (another work in progress).

 

The talk is intended for a general audience and will be given in Hebrew (with an English presentation).

 

 

ההרצאה המיועדת לקהל רחב תעסוק בשני קווים נפרדים בהם עסקתי לאורך 40 השנה האחרונות. מה המשמעות של סטטיסטיקה בייסיינית בתקופת הנתונים הגדולים? אטען כי הסטטיסטיקה הבייסיינית המבוססת על פריור המבטא אמונה אינה עומדת בדרך הכלל בכללים הלא בייסייניים. אסיים פרק זה עם מאמר עכשווי על מינימקס במודל של החלטות משולבות. הקו השני ידון במודל הרגרסיה החלקית לינארית (PLM, דהיינו מודל ריגרסיה המכיל מרכיב לינארי ומרכיב לא פרמטרי), כמודל סמי-פרמטרי, ומקבילו היותר מודרני, היסק סטטיסטי במודל בעלי מספר עצום של משתנים מסבירים. אסיים במאמר עכשווי המציג את מודל הריגרסיה עם חוסר רציפות (RDD) כמודל מסוג זה. מודל זה האחרון מתאים לדוגמא, לאמידת האפקט של קבלה למסלול לימודים יוקרתי על עתיד התלמידים.

ההרצאה תנתן בעברית עם שקפים בלע"ז. 

 

Location: 

Eilan Hall, Feldman Building, Second Floor, Edmond Safra Campus.

Click here to add the "In Conversation" Series to your Google Calendar