zoukankan      html  css  js  c++  java
  • The Jensen Inequality

    Let(设) f be a function with a positive second derivative(二阶导数). Such a function is called “convex"(凸的,注意是向下凸,向上凹,如 x2 ) and satisfies the inequality  

      \begin{displaymath}
{f(a)\ +\ f(b)\over 2}\ -\ f\left( {a + b\over 2}\right)\quad \geq \quad 0\end{displaymath}                      (1)

    inequation (1) relates a function of an average to an average of the function. The average can be weighted, for example,  

      \begin{displaymath}
{ \frac{1}{3} \, f(a)\ +\ 
 \frac{2}{3} \, f(b)}\ -\ f\left( {
 \frac{1}{3} a +
 \frac{2}{3} b}\right)
\quad \geq \quad 0\end{displaymath}            (2)

    Figure 1 is a graphical interpretation of inequation (2) for the function f=x2

    jen

    There is nothing special about f=x2, except that it is convex. Given three numbers a, b, and c, the inequality (2) can first be applied to a and b, and then to c and the average of a and b. Thus, recursively, an inequality like (2) can be built for a   weighted average(加权平均数)    of three or more numbers. Define weights $w_j \geq 0$ that are normalized(标准化) ($\textstyle {\sum_j} w_j = 1$). The general result(通式) is  

      \begin{displaymath}
S(p_j) \eq \sum_{j=1}^N w_j f(p_j)\ -\ f\left(
 \sum_{j=1}^N w_j p_j \right) \quad \geq \quad 0\end{displaymath}    (3)

    If all the pj are the same, then both of the two terms in S are the same, and S vanishes. Hence, minimizing S is like urging all the pj to be identical(完全一样). Equilibrium is when S is reduced to the smallest possible value which satisfies any constraints that may be applicable. The function S defined by (3) is like the entropy(熵) defined in thermodynamics(热力学).

  • 相关阅读:
    CentOS7下Elastic Stack 5.0日志分析系统搭建
    ElasticSearch 简单入门
    简单使用packetbeat
    centos7没有安装ifconfig命令的解决方法
    CentOS系统下docker的安装与卸载
    centos7 cannot find a valid baseurl for repo base
    HP P2xxx/MSA SMI-S Provider
    Zookeeper 的学习与运用
    kafka入门:简介、使用场景、设计原理、主要配置及集群搭建(转)
    利用开源架构ELK构建分布式日志系统
  • 原文地址:https://www.cnblogs.com/kevinGaoblog/p/2425986.html
Copyright © 2011-2022 走看看