Inverse Regression in Binary Response LDV Model |
| |
Authors: | Jianjun Ma Xingzhong Xu Junlong Zhao |
| |
Institution: | 1. Department of Mathematics , Beijing Institute of Technology , Beijing, P.R. China;2. Department of Mathematics , Shenyang Institute of Technology , Shenyang, P.R. China jianjunmark2006@163.com;4. Department of Mathematics , Beijing Institute of Technology , Beijing, P.R. China |
| |
Abstract: | Sliced inverse regression, a link-free and distribution-free method, is applied to binary response limited dependent variable models. An inverse regression property of binary response LDV model is found. Based on this property, if the distributions of X j (j = 1, 2,…, p) satisfy the linearity condition, then β can be estimated up to a positive multiplicative scalar without any assumptions on the distribution of error ε. Moreover, the estimator can be proved to be asymptotically normal based on which testing hypotheses are considered. Simulations results are reported. |
| |
Keywords: | Binary response model δ method Limited dependent variable models Sliced inverse regression |
|
|