Figure 5:

The choice of ε in the estimation of $\mathbb{E}\left[Y_i\left(t;x_i,x_j,w_j\right)\right]$ with sample size n = 100, 000 under the constant hazards α(t) = 0.2, γ(t) = 10 and coefficients $e^{{\beta }_0}=e^{{\beta }_1}=0.2$ and e^σ = 0.5. Figure on the left shows the estimation of $\mathbb{E}\left[Y_i\left(t=2;x_i=1,x_j=1,s=1\right)\right]$ and its corresponding MSE under different choices of ϵ, and Figure on the right shows the estimation of $\mathbb{E}[Y_i(t=1.5;x_i=0,x_j=0,s=1)]$ and its corresponding MSE under difference choices of ϵ.