- Hands-On Mathematics for Deep Learning
- Jay Dawani
- 269字
- 2021-06-18 18:55:25
Multiple random variables
A lot of the time, we will end up dealing with more than one random variable. When we do have two or more variables, we can inspect the linear relationships between the random variables. We call this the covariance.
If we have two random variables, X and Y, then the covariance is defined as follows:

The following are some of the axioms for the covariance:
- If c is a constant, then
.
.
.
.
.
.
, given that X and Y are independent (but it does not imply that the two are independent).
However, sometimes, the covariance doesn't give us the full picture of the correlation between two variables. This could be a result of the variance of X and Y. For this reason, we normalize the covariance as follows and get the correlation:

The resulting value will always lie in the [-1, 1] interval.
This leads us to the concept of conditional distributions, where we have two random variables, X and Y, that are not independent and we have the joint distribution, , from which we can get the probabilities,
and
. Then, our distribution is defined as follows:

From this definition, we can find our conditional distribution of X given Y to be as follows:

We may also want to find the conditional expectation of X given Y, which is as follows:

Now, if our random variables are independent, then, , which we know to be true because Y has no effect on X.
- 數據庫原理及應用教程(第4版)(微課版)
- Access 2016數據庫教程(微課版·第2版)
- 劍破冰山:Oracle開發藝術
- Voice Application Development for Android
- Effective Amazon Machine Learning
- 大數據時代下的智能轉型進程精選(套裝共10冊)
- MATLAB Graphics and Data Visualization Cookbook
- 深入淺出 Hyperscan:高性能正則表達式算法原理與設計
- Solaris操作系統原理實驗教程
- Visual Studio 2013 and .NET 4.5 Expert Cookbook
- Expert Python Programming(Third Edition)
- Access數據庫開發從入門到精通
- MySQL數據庫應用與管理
- Access 2010數據庫應用技術教程(第二版)
- Kafka權威指南(第2版)