# week1
## 数量积
（又称点积）是输入两个向量，输出实数值标量的二元运算，是欧几里得空间的标准内积
ab = a1b1+a2b2+...
使用矩阵乘法内积可以表示为 a.b = aT*b =|a||b|cos
# week2
## 感知器学习算法PLA
```math
\sum_{i = 1}^{n}(w_ix_i) > t
\sum_{i = 1}^{n}(w_ix_i - t) > 0
\sum_{i = 1}^{n}(w_ix_i + w_0x_0) > 0 \quad (w_0= -t,x_0 = 1)
\sum_{i = 0}^{n}(w_ix_i) > 0
w^Tx > 0 =>y=1
w^Tx < 0 =>y=-1
sign(w^Tx) = y \quad (w:weight\ vector, x:feature\ vector)
--------------
for \ t = 0,1,2...n
if\ sign(w_t^Tx_{k(t)}) \neq y_{k(t)}
w_{t+1} = w_t +y_{k(t)}.x_{k(t)}
```
原文：http://www.cnblogs.com/dailing/p/7764442.html