Table 1: Examples of ρ and ψ functions used with M-estimators.
P( t ) |
___________ψ ( t )___________ | |
Least squares |
t2 |
21 |
Least absolute deviation |
∖t∖ |
sign(t) |
Quantile estimation |
{τ — I(x < 0) }x |
τ — I(x < 0) |
Huber: for ∖t∖ ≤ c |
t2 |
2t |
for c < ∖t∖ |
c∖t∖ |
c sign(t) |
Hampel: for ∖t∖ ≤ a |
t2 |
2t |
for a < ∖t∖ ≤ b |
a∖t∖ |
a sign(t) |
for b < ∖t∖ ≤ c |
ɪt — -∖t2 sign(t) |
a ( c — ∖t∖ ) /( c — b ) |
for c < ∖t∖ |
a∖t∖ |
0 |
Biweight (Tukey) |
— (c2 — t2)31(∖t∖ ≤ c)/6 |
t(c2 — t2)21(∖t∖ ≤ c) |
Sine (Andrews) |
—ccos(x/c)I(∖t∖ ≤ πc) |
sin(x/c)I(∖t∖ ≤ πc) |
10
More intriguing information
1. Globalization, Redistribution, and the Composition of Public Education Expenditures2. Heterogeneity of Investors and Asset Pricing in a Risk-Value World
3. Micro-strategies of Contextualization Cross-national Transfer of Socially Responsible Investment
4. SOME ISSUES IN LAND TENURE, OWNERSHIP AND CONTROL IN DISPERSED VS. CONCENTRATED AGRICULTURE
5. Trade Openness and Volatility
6. Natural Resources: Curse or Blessing?
7. Staying on the Dole
8. The storage and use of newborn babies’ blood spot cards: a public consultation
9. An Investigation of transience upon mothers of primary-aged children and their school
10. The name is absent