Content based filtering

Bayes Classifier

Background knowledge:

By comparing values of $P(c(\bar{X})=1|x_1,...,x_d)$ and $P(c(\bar{X})=-1|x_1,...,x_d)$, one can determine whether or not the active user likes $\bar{X}$. To predict $P(c(\bar{X})=1)$, the following expression is used.

$$P(c(\bar{X})=1)=\dfrac{|D_L^+|+\alpha}{|D_L|+2\cdot \alpha}$$

In this expression, $$

$$f(x) = x * e^{2 pi i \xi x}$$