From af55e7c9c99f48b3a0a9b28c6b81fb8314c7f787 Mon Sep 17 00:00:00 2001
From: Kailing Ding <markdingkl@gmail.com>
Date: Tue, 18 Jan 2022 16:19:38 -0800
Subject: [PATCH] add examples fors quantile regression

---
 .../quantile-regression.md                    | 28 +++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/website/docs/uncertainty-quantification-methods/quantile-regression.md b/website/docs/uncertainty-quantification-methods/quantile-regression.md
index 02863eda..9ed2d2de 100644
--- a/website/docs/uncertainty-quantification-methods/quantile-regression.md
+++ b/website/docs/uncertainty-quantification-methods/quantile-regression.md
@@ -10,3 +10,31 @@ L_{Quantile}\big(y,f(x),\theta,p\big) = min_\theta\{\mathbb{E}_{(x,y)\sim D}[(y
 $$
 
 where $p$ is our fixed confidence interval parameterized by $\theta$. When the pinball loss is minimized, the result is the optimal quantile.
+
+### Examples
+
+---
+
+```python
+from torchts.nn.loss import quantile_loss
+from torchts.nn.model import TimeSeriesModel
+
+# initialize model
+class Model(TimeSeriesModel):
+    def __init__(self):
+        ...
+
+    def forward(self, x):
+        ...
+
+model = Model(
+    ...,
+    criterion=quantile_loss,
+    criterion_args={"quantile": 0.05}
+)
+
+model.fit(...)
+y_pred = model.predict(...)
+```
+
+[Full example](https://github.com/Rose-STL-Lab/torchTS/blob/main/examples/quantile-regression/lstm-quantile-regression.ipynb)