Analysis of the physicochemical properties, bacteria
numbers, and sensory quality of vacuum-packaged compressed
biscuits (VCB) was conducted at room temperature
and (or) elevated temperatures to ascertain the main deterioration
indicator. The acid value (AV) was determined as the
deterioration index and the limiting value was 3.28 mg g−1.
The deterioration curves of VCB between 45 and 85 °C were
investigated and the changes of AV followed a first-order
reaction. A modified Arrhenius model was selected for shelf
life prediction (SLP). For analyzing the factors associated with
prediction accuracy, the R2, %RMS values of the prediction
model, and the prediction errors of shelf life from accelerated
shelf life tests (ASLT) were evaluated. The results showed that
the number of replicates was the most important factor for
prediction accuracy, followed by the time intervals for sampling
and the number of points fitted for determining the
deterioration curve. At the appropriate sampling interval, 9
points fitted in the AV change curves with three replicates
were necessary. In the experimental design of ASLT, the
results indicated that the number of temperatures had the
greatest impact on the accuracy and at least one low temperature
close to room temperature should be used when designing ASLT. However, for a long shelf life product such
as VCB which the shelf life is more than 2 years, on the
premise of prediction accuracy, taking into account the test
cycle and cost, a certain number of relative higher temperatures
(65, 75, and 85 °C) can be used in ASLT.