Intrinsic Dimension Estimation-Based Feature Selection and Multinomial Logistic Regression for Classification of Bearing Faults Using Compressively Sampled Vibration Signals

. 2022 Apr 5;24(4):511. doi: 10.3390/e24040511

Algorithm 1 Vibration data compression using CS

1. Input: vibration dataset

X \in R^{n x L};

measurement matrix

Ө \in R^{m x n}

and compressive sampling rate

α

2. Output: compressively sampled vibration signals

Y \in R^{m x L}

3. Produce the sparse representations

S

of X:

a b s (F F T (X))

\to S \in R^{n x L}

4. Project

S

into

Ө

with compressed sampling rate

α

to obtain compressively sampled to obtain compressively sampled signals

Y \in R^{m x L}