this is for holding javascript data
Xavier Holt added We_then_substitute_our_estimated__.tex
over 8 years ago
Commit id: 5f57f0a3eed0b864bd2864ec6f6e0d7fa7ab2443
deletions | additions
diff --git a/We_then_substitute_our_estimated__.tex b/We_then_substitute_our_estimated__.tex
new file mode 100644
index 0000000..de4ccbd
--- /dev/null
+++ b/We_then_substitute_our_estimated__.tex
...
We then substitute our estimated weight-distribution back into the integral above, allowing us to predict new points. In terms of implementation, in this paper we use the Laplacian prior distribution with $\lambda_i = \lambda \ \forall i$ being an equal regularisation hyperparameter that for our initial experiments we simply set to 10 through cross-validation and inspection.