https://github.com/sherif-mooo/mts-cpca
Common PCA for Multivariate time series
https://github.com/sherif-mooo/mts-cpca
multivariate-time-series pca-analysis
Last synced: 4 days ago
JSON representation
Common PCA for Multivariate time series
- Host: GitHub
- URL: https://github.com/sherif-mooo/mts-cpca
- Owner: Sherif-MoOo
- Created: 2022-01-08T19:37:46.000Z (over 4 years ago)
- Default Branch: main
- Last Pushed: 2022-02-23T06:57:48.000Z (over 4 years ago)
- Last Synced: 2025-02-23T18:15:31.939Z (over 1 year ago)
- Topics: multivariate-time-series, pca-analysis
- Language: Python
- Homepage:
- Size: 184 KB
- Stars: 2
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# PCA-FOR-MTS
I couldn't find suitable framework that deals with MTS-CPCA so I'll build the algorithm from scratch
What's MTS?
Time series data can be seen everywhere including stock exchange, financial market, medicine and engineering,
which is one kind of the important data needed to be mined for the valuable information and knowledge. It has two
categories according to various number of the variables, they are univariate time series (UTS) and multivariate time
series(MTS).
MTS (Multivariate time series) is an important type of data that
is indispensable in a variety of domains as medicine
domain which is the evolution of a group of synchronous
variables over a duration of time as shown in the figure and
there is a lot of effort given due to the expensive of gathering
these labeled data to be able to offer a method gives a reliable
accuracy by only using a limited amount of these data.
What's PCA?
This is a linear unsupervised algorithm to find orthogonal transformation axes that diagonalize the covariance
matrix the goal is to eliminate low variance and highly correlated features.
Why PCA?
Due to the high
dimensionality of MTS, the dimensionality reduction is proposed to validly integrate into the clustering,classification and regression process and a good MTS accuracy can be obtained in lower dimensions
Suppose there was a dataset X having N multivariate time series
Σ_i=cov(x_i) , x_i ∈ R^(n_i * m) where n_i is the length of MTS sample and m is the number of the variables
Σ = 1/N ∑Σ_i