알파트로스

[AFML] 2.3.1 Financial Data Structures - Standard Bars 본문

Advances in Financial Machine Learning

[AFML] 2.3.1 Financial Data Structures - Standard Bars

알파트로스 2021. 7. 17. 02:31
In order to apply ML algorithms on your unstructured data, we need to parse it, extract valuable information from it, and store those extractions in a regularized format.

 

2.3.1 Standard Bars

The purpose of these method is to transform a series of observations that arrive at irregular frequency (often referred to as 'inhomogeneous series') into a homogeneous series dereived from regular sampling

 

2.3.1.1 Time Bars

  • sampling information at fixed time intervals
  • most popular but they should be avoided
  1. markets do not process information at a constant time interval
    > the hour following the open is much more active than the hour around noon 
    > oversample information during low-activity periods
  2. exhibit poor statistical properties, like serial correlation, heteroscedasticity and non-normality of returns
코스피200 Time Bars 예시 

매일 모이는 데이터의 가장 첫행과 마지막 행은, 동시호가로 정보가 뭉쳐있어 제거해 주었다

2.3.1.2 Tick Bars

  • extracted each time pre-defined number of transactions take place. this allows us to synchronize sampling with a proxy of information arrival
  • many studies have confirmed that sampling as a function of trading activity allows us to achieve returns closer to IID Normal : This is important, because many statistical methods rely on the assumption that obsevations are drawn from an IID Gaussian process. Intuitively, we can only draw inference from a random variable that is invariant 
  • when constructing tick bars, need to be aware of outliers. Many exchanges carry out an auction at the open and an auction at the close

2.3.1.3 Volume Bars

  • sampling every time a pre definded amount of the security's units have been exchanged
  • volume achieved even better statistical properties than sampling by tick bars
  • several market microstructure theories study the interaction between prices and volumes

2.3.1.4 Dollar Bars

  • sampling an observation every time a pre-definded market value is exchanged 
  • robust in the face of corporate actions 

#  틱 바 볼륨바 달러바의 평균 갯수 그려보기 정리하기