Biomedical Engineering Reference
In-Depth Information
Individual
records
Multiple
records
Linear
relationships
FIGURE 5.5
:
Random data analysis
“Correlation Function” using either Time or Frequency Domain approach. Periodic-
ity may also be analyzed via the “Power Spectral Analysis.” The next step regarding
random data may be to examine the distribution of the data (“Is the data Normality
distributed?”) or to determine all the basic statistical properties (Moments) of the signal
(data).
5.5 RANDOM DATA ANALYSIS
The fifth and final step is data analysis of the data supplied as time histories of some
physical phenomenon. Data analysis involves three distinct and independent processes
dependent on the experimental question and the data set as shown in Fig. 5.5:
1.
analysis of individual records,
2.
analysis of multiple records, and
3.
linear relationships.
If the analysis is to be accomplished on an individual record, that is, one chan-
nel of electrocardiogram recording, or one channel of electroencephalogram recording,
etc., then either “Autocorrelation Function Analysis” (Time-domain Analysis) or “Au-
tospectral Analysis” (Frequency-domain Analysis) would be preformed on the random
data.
If the analysis is to be accomplished on multiple records, that is, more than one
channel of recording, then either “Cross-correlation Function Analysis” (Time-domain
Analysis) or “Cross-spectral Analysis” (Frequency-domain Analysis) would be preformed
on the random data. Multiple Records would be used to determine the “Transfer
Functions” of systems when the input stimulus and the output responses are random
signals.
Search WWH ::
Custom Search