0

Suppose I have a (training) set of $n$ observation $\{(Y_i^{(U)},X_i^{(U)})\}_{i=1}^n$ of age $X_i^{(U)}$ and height $Y_i^{(U)}$ from people in the USA.

Now suppose I also have a (test) set of $m$ observations $\{X_i^{(J)}\}_{i=1}^m$ of age $X_i^{(J)}$ only from people in Japan, where people are shorter on average.

I want to predict the heights of people in Japan in the test set using transfer learning from the USA dataset. Suppose for simplicity the USA data is well-fit by the standard simple linear regression model $$ Y_i^{(U)} = \beta_0 + \beta_1 X_i^{(U)} + \varepsilon_i, \quad \quad i=1,\dots,n, $$ with the usual simple linear regression assumptions for the error.

How can I use transfer learning of the USA data to make predictions of height given the age data in the Japan dataset $\{X_i^{(J)}\}_{i=1}^m$?

Are there any theoretical results on the accuracy of the predictions given by transfer learning?

sonicboom
  • 101
  • 1
  • 1
    Perhaps use the same model but adjust the predictions by adding an offset representing the difference in average heights between Japanese and US people. – Jayaram Iyer May 02 '21 at 07:32
  • 1
    Transfer learning in the strict sense is applied to deep learning models, not linear regression, but possibly addnin some extra parameters in your regression model can accoutn for the discrepancies – Nikos M. May 03 '21 at 09:37

0 Answers0