||
An Application of System Kinematics on a Stock Market
Definition of 30-market: Thirty stocks were chosen from Fidelity, and these thirty stocks in a given order make up the 30-market.
Definition of 30-vectors: The thirty prices of these thirty stocks expresses the state of the 30-market.
Y=Y(i) (i=1,2,…30), (1-1)
Y’=Y(i)/|Y| (i=1,2,…30), (1-2)
Where the |Y|=√∑(( Y(i))2), (i=1,2,…30),is the vector length, a scalar.
The market transition vectors (trends) are defined as the present over the previous states:
T(i,k)=Y’(i,k)/Y’(i,k-1) (i=1,2,…30) (2-1)
Where the second index indicates the time: k=present, k-1=previous.
Projection of the next states are expressed by a vector product of the present states and the trends:
P(i,k+1)=Y(i,k)*T(i,k) (i=1,2,…30) (2-2)
Where the P(i,k+1) is the projection of the next states of the 30-market based on the existing information of the 30-market.
Bai-Jameson Filter connects the two sequential states:
Y(i,k+1)=[Y(i,k)*T(i,k)* β +D(i,k+1)* α]/(α+β) (i=1,2,…30) (3-1)
Where the D(i,k+1) are the 30-observation vectors, the closing prices of the thirty stocks on time k+1. Thus, the expectation of the 30-market on k+1 are the weighted sum of the observations on k+1 and the projection from k. And, where α and βare the weights of D and P, respectively. Setting the two weights equal to α = β =0.5:
Y(i,k+1)=[Y(i,k)*T(i,k)+D(i,k+1)]/2 (i=1,2,…30) (3-2)
The formula can be further extended into a time chain (3-3). This Bai-Jameson Chain uses history information to predict the future states in a manner of Fading Memory:
P(i,k+1)= T(i,k)(D(i,k)+0.5*T(i,k-1)(D(i,k-1)+0.5*T(i,k-2)(D(i,k-2)+...+0.5*T2(D2+D1))))
(i=1,2,…30) (3-3)
Main Reference: T. Jay Bai, Trend Analysis, The Ethnic Publishing House, Beijing, 2006.
Translated from:http://blog.sciencenet.cn/blog-333331-768774.html
“新‘系统论’初探”地址:http://blog.sciencenet.cn/blog-333331-825601.html
Archiver|手机版|科学网 ( 京ICP备07017567号-12 )
GMT+8, 2024-11-24 06:56
Powered by ScienceNet.cn
Copyright © 2007- 中国科学报社