之所以OUT
是因為沒看延禧攻略
在統計學中，我們需要研究兩個變量之間的關係，然後可以推算出兩個變量之間關係的model模型表達式，進而可以進行預測估算。
但是對於兩個變量之間的具體關係，其實常常會考到一個知識點。
那就是correlation and causation的區分。
第一點：
Correlation shows how closely two variable vary with each other. For example, as the value of one increases, so does the other.
兩個變量之間只是數量上面存在某種關係。
Causation is when two variables directly affect each other.
For example, the time you go to bed affects the number of hours you sleep.
兩個變量之間會相互影響，存在某種因果關係。
Sometimes a cause and effect are closely related, but not always. It is easy to assume that events that are closely correlated are also connected causally. But correlation between two events does not mean that one has caused the other.（有相關性並不代表有因果關係）
第二點：
那麼相關性和因果關係到底怎麼確定呢，
有些可以根據生活常識或者科學原理去得到結論，
但是大部分情況其實我們的主觀意識不一定是正確的，數據也不能單方面的得到因果關係，
比如，氣溫的高低和股票指數也許湊巧就有正相關或者負相關的關係，
但是兩者之間根本沒有任何關係，除了數量上看起來的湊巧關係，
因為有可能有其他隱形要素在影響，
這時候就需要藉助experiment試驗的嚴密設計，才能真正去驗證。
「CORRELATION NEVER PROVE CAUSITION」
Experimental research investigates what happens when you change a variable, for example what happens to a liquid when you increase the temperature.
Correlated research does not change the variables. It observes the outcome of two events and offers statistical data as proof.
我沒看延禧攻略，那就out了吧汪汪汪
閱讀更多 MathsAlgorithm 的文章
