3.3. Statistical analysis of electronic nose sensor
response data
In order to observe whether the sensor array was able to distinguish different ripeness states, PCA and LDA analyses were
applied to the 150 measurements (30 samples for each picking
date) performed with the olfactory system and results were
shown inFig. 4. By PCA, the samples in group of week 1 were
clustered on the right of the plot, the samples of the week 2
samples were located in the middle of the plot,Fig. 4a shows a
partial overlapping between week 3 group and week 4 group,
between week 4 group and week 5 group. After analyzing the
same data set using LDA, five groups (Fig. 4b) were distinguished. In this plot, about 96% of the total variance of the
data is displayed. Functions 1 (LD1) and 2 (LD2) accounted for
68% and 28% of the variance, respectively. A clear evolution in
ripening process can be observed.
3.3. Statistical analysis of electronic nose sensorresponse dataIn order to observe whether the sensor array was able to distinguish different ripeness states, PCA and LDA analyses wereapplied to the 150 measurements (30 samples for each pickingdate) performed with the olfactory system and results wereshown inFig. 4. By PCA, the samples in group of week 1 wereclustered on the right of the plot, the samples of the week 2samples were located in the middle of the plot,Fig. 4a shows apartial overlapping between week 3 group and week 4 group,between week 4 group and week 5 group. After analyzing thesame data set using LDA, five groups (Fig. 4b) were distinguished. In this plot, about 96% of the total variance of thedata is displayed. Functions 1 (LD1) and 2 (LD2) accounted for68% and 28% of the variance, respectively. A clear evolution inripening process can be observed.
การแปล กรุณารอสักครู่..
