Data Enlargement

資料的擴展

凌嘉華(Chiahua, Ling)

《致理學報》 40期 (2020/11) Pp. 1-9

摘要

The purpose of this paper is to make a new way which imitates the idea of Bayesian approach with a difference. In Bayesian approach one considers a statistical prior distribution which is centered at super parameters, and we assume the way ( data increase) is to improve for small samples where a very few observations (data) are available. And using statistical prior distributions, at the core of the given observations ( calling them to be hyper-parameters ), we may call as second generation dataset by producing a bigger dataset, then we can make statistical inferences by using this bigger second dataset.

關鍵字

Bayesian method ； Conjugate distribution ； statistical prior distribution

並列摘要

這篇文章用和貝氏方法有些微不同，貝氏技術有先驗分配和以超參數為中心，然後我們增大資料先採用小樣本裡較少的觀測值來做基礎，用以觀測值（假設它們是超參數）做為核心的統計先驗分配，之後產生一個比較大的資料群體（我們稱做第二個產生資料群體），第二個產生的資料群體則是用來做統計的推論和資料分析。

並列關鍵字

貝氏方法 ； 共軛分配 ； 統計先驗分布