adjust a typo

Former-commit-id: 641c95a9780050251a9188b2a2f3e7e65d19b2ec

Chapter13/linear_factor_models.tex

@@ -227,22 +227,24 @@ \section{\glsentrytext{SFA}}
 % 484
 
 
-\glssymbol{SFA}算法~\citep{WisSej2002}包括将$f(\Vx;\theta)$定义为线性变换，并求解满足如下约束
+
+\glssymbol{SFA}算法~\citep{WisSej2002}包括将$f(\Vx;\theta)$定义为线性变换，并求解如下优化问题
 \begin{align}
-\label{eqn:139}
-\SetE_t  f(\Vx^{(t)})_i = 0 
+	\label{eqn:138}
+	\min_{\Vtheta} \SetE_t  (f(\Vx^{(t+1)})_i - f(\Vx^{(t)})_i  )^2
 \end{align}
-以及
+并且满足下面的约束：
 \begin{align}
-\label{eqn:1310}
-\SetE_t [ f(\Vx^{(t)})_i^2 ] =1 
+	\label{eqn:139}
+	\SetE_t  f(\Vx^{(t)})_i = 0 
 \end{align}
-的优化问题
+以及
 \begin{align}
-\label{eqn:138}
-\min_{\Vtheta} \SetE_t  (f(\Vx^{(t+1)})_i - f(\Vx^{(t)})_i  )^2.
+	\label{eqn:1310}
+	\SetE_t [ f(\Vx^{(t)})_i^2 ] =1. 
 \end{align}
 
+
 % 485
 
 学习特征具有零均值的约束对于使问题具有唯一解是必要的;