diff --git a/markdown/week8.md b/markdown/week8.md
index 5b607b47..b48d2c9d 100644
--- a/markdown/week8.md
+++ b/markdown/week8.md
@@ -327,7 +327,7 @@ $$z^{(i)}=U^{T}_{reduce}*x^{(i)}$$
 ![](../images/a4477d787f876ae4e72cb416a2cb0b8a.jpg)
 
 其中的$S$是一个$n×n$的矩阵，只有对角线上有值，而其它单元都是0，我们可以使用这个矩阵来计算平均均方误差与训练集方差的比例：
-$$\dfrac {\dfrac {1}{m}\sum^{m}_{i=1}\left\| x^{\left( i\right) }-x^{\left( i\right) }_{approx}\right\| ^{2}}{\dfrac {1}{m}\sum^{m}_{i=1}\left\| x^{(i)}\right\| ^{2}}=1-\dfrac {\Sigma^{m}_{i=1}S_{ii}}{\Sigma^{k}_{i=1}S_{ii}}\leq 1\%​$$
+$$\dfrac {\dfrac {1}{m}\sum^{m}_{i=1}\left\| x^{\left( i\right) }-x^{\left( i\right) }_{approx}\right\| ^{2}}{\dfrac {1}{m}\sum^{m}_{i=1}\left\| x^{(i)}\right\| ^{2}}=1-\dfrac {\Sigma^{k}_{i=1}S_{ii}}{\Sigma^{m}_{i=1}S_{ii}}\leq 1\%$$
 
 也就是：$$\frac {\Sigma^{k}_{i=1}s_{ii}}{\Sigma^{n}_{i=1}s_{ii}}\geq0.99$$