In recent years, numerous attempts have been made to reduce the sugar content of processed foods to meet consumer demand. The aim of this study was to produce cookies without added sugar through rice koji fermentation combined with the addition of food enzymes. Rice koji prepared from the yukikomachi variety, which exhibits high enzyme activity, was the most suitable among five tested white koji molds for cookie production. The thickness and expansion rates of cookies prepared using rice koji and food enzymes were lower than those of sugar-containing cookies (control). However, no significant difference in cookie diameter was observed, indicating a greater spread in experimental cookies. Compared to the control, the L* values for the cookies were significantly lower, whereas the a* and b* values were higher, consistent with increased browning. Experimental cookies were markedly softer than the control cookies, and brittleness differed significantly among treatments. Differences in sugar composition were also observed. Further, total free amino acid and GABA contents were approximately 20.5–25.8-fold and 1.4–1.7-fold higher than the control, respectively. Additionally, the experimental cookies had fewer calories, higher crude protein levels, and lower carbohydrate levels compared to the control cookies. These results suggest that cookies prepared with rice koji and food enzymes may serve as a healthier alternative.

We developed a forecasting model for estimating the temperature of fresh-cut Japanese radish samples placed in sealed plastic containers and subjected to rapid changes in temperature during storage. We constructed the forecasting model using an autoencoder and deep neural network, applying time series data of temperature in both incubators and plastic containers to predict the temperature time series data in the samples. We found good agreement between the predicted and observed temperature profiles, indicating that our forecasting model could also be used to predict the accumulated storage temperature (AST) of samples. We further tested the applicability of the forecasting model under different types of rapid change experiments. The results showed that the model was able to reproduce AST as well as the temperature time series in samples stored in sealed plastic containers. We expect that our new model will contribute to innovative technologies for predicting the temperature of fresh foods in locations where direct measurement is not possible.